Kidney International, Vol. 48 (1995), pp.

469—474

Epidemic aluminum intoxication in hemodialysis patients traced

to use of an aluminum pump
DALE R. BURWEN, STEVEN M. OLSEN, LEE A. BND, MATFHEW J. ARDUINO, MARIE H. REID,
and WILLIAM R. JARvIs

Hospital Infections Program, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Public Health Service, U.S. Department
of Health and Human Services, Atlanta, Georgia, and Center for Devices and Radiological Health, Food and Drng Administration, Public Health
Service, US. Department of Health and Human Services, Rockville, Matyland, USA

Epidemic aluminum intoxication in hemodialysis patients traced to use in 1981 [21]. Two other sources of epidemic aluminum intoxica-
of an aluminum pump. This study was designed to identify the source, risk tion in dialysis patients have been reported: a boiler aluminum
factors, and clinical consequences of an outbreak of aluminum intoxica-
tion in hemodialysis patients using case-control and cohort studies, In anode that contaminated water used in dialysate [22], and an
1991, a dialysis center in Pennsylvania [Dialysis Center A (DCA)] aluminum-containing cartridge used in a sorbent regeneration
identified a number of patients with elevated serum aluminum levels. All system, which disintegrated and contaminated the dialysate [23].
patients receiving dialysis at DCA during Januaiy 1, 1987 to March 26, Currently, although aluminum intoxication in renal failure
1992 were involved in the study. A case-patient was defined as any patient patients occurs sporadically due to oral aluminum-containing
with a serum aluminum level 100 sgIliter after dialysis sessions at
DCA. Fifty-nine case-patients were identified. Risk factors for elevated medications [24—28], reports of the epidemic occurrence of
serum aluminum levels were receipt of bicarbonate- (rather than acetate-) aluminum intoxication in dialysis centers, with associated severe
based dialysate, higher number of sessions using bicarbonate dialysis, clinical manifestations, have virtually disappeared.
receipt of acid concentrate (used in bicarbonate dialysis) passed through However, in 1991, a dialysis center in Pennsylvania [Dialysis
one of two electric pumps, and a greater number of sessions using this
concentrate. The electric pumps had an aluminum casing, casing cover, Center A (DCA)], identified a number of patients with elevated
and impeller. Elevated levels of aluminum were found in acid concentrate serum aluminum levels. In 1988, structural renovations and the
after passing through a pump. Seizures and mental status changes installation of electric pumps to deliver concentrates used in
requiring hospitalization were associated with aluminum exposure. We dialysate from the ground floor to the second floor had occurred.
found that epidemic aluminum intoxication was caused by the use of an From 1988 through 1992, bicarbonate dialysis use had increased.
electric pump with aluminum housing to deliver acid concentrate used in
bicarbonate dialysis. This outbreak demonstrates why it is essential to When the initial investigation suggested contaminated dialysate,
insure that all fluid pathways, storage tanks, central delivery systems, and DCA discontinued use of the delivery system, and the Centers for
pumps are compatible with low pH fluids before converting from acetate Disease Control and Prevention (CDC) and the Food and Drug
to bicarbonate dialysis. Administration were notified and assisted in the investigation.
Methods
Aluminum intoxication can cause dialysis encephalopathy, also Identification of patients with elevated serum aluminum levels
called dialysis dementia [1], a demineralizing osteodystrophy [2], To identify patients with elevated serum aluminum levels, we
and anemia [3]. The two severe manifestations, dialysis encepha- reviewed medical records of all DCA patients present during
lopathy and fracturing osteodystrophy, were reported as epidemic January 1, 1987 to March 26, 1992 (study period). Most patients
occurrences in hemodialysis centers worldwide from the late had at least one, and commonly had quarterly, aluminum levels.
1960's through the early 1980's [4—20]. The source of these Monthly aluminum testing was started in January 1992.
outbreaks was usually parenteral exposure to high concentrations
of aluminum in water used in dialysate in areas where alum or Case/noncase definitions
aluminum sulfate was added as a flocculent to municipal water Next, we defined a case-patient as any patient during the study
supplies. To address this problem, water treatment with reverse period with a serum aluminum level 100 .rg/liter (normal 0 to
osmosis units or deionizers was shown to remove aluminum and 12) after 5 dialysis sessions at DCA. Aluminum levels measured
other harmful chemicals [12, 13, 17, 18], and the Association for after administration of deferoxamine, which can increase serum
the Advancement of Medical Instrumentation (AAMI) published aluminum levels, were excluded. We defined a control or noncase-
standards for maximum aluminum levels in water used in dialysate patient as any patient during the study period with a maximum
serum aluminum level <100 pg/liter after 5 dialysis sessions at
DCA.
Received for publication October 14, 1994
and in revised form February 1, 1995 Analysis of risk factors for elevated aluminum levels
Accepted for publication March 9, 1995 Since the majority of case-patients had elevated aluminum
© 1995 by the International Society of Nephrology levels after October 15, 1988, the approximate date of installation

469
470 Burwen et al: Epidemic aluminum intoxication

20
19 -
18
17
16
15
14
Cl)13 -
C
a) 12 -
0 11 -
Ca

0 10 -
9-
a)
.0
E 8-
z 7-

I
6
A
5
4
3
2
1

0
C.) .00.- C C)
I
nRfli
I

, a) .0
I

- C C)
I I I

0
a) .0 - C D)
—
C 0) 0 C.) .0
L ¶7<9C L<?< OO9?< 9c a) 0. '- cI 0.
C.)
?i. Q)
0)

0.> 7<
>'- 9OLL
0.
(1)

a) 0
C,) z -, a) 0 Ca
1987 1988 1989 1990 1991 1992
Date of first serum aluminum level 100 tg/Iiter
Fig. 1. Distribution of patients with semm aluminum level 100 pg/liter by month, DCA, January 1, 1987 to March 26, 1992. Sixty-four patients had a serum
aluminum level 100 pg/liter during the study period; one patient with first aluminum level pg/liter before 1987 is not shown. In January 1992,
a large number of patients with a serum aluminum level 100 gg/liter were identified due to a point prevalence survey. At A, two electric pumps were
installed at DCA, approximately October 15, 1988; at B, acid concentrate was initiated through one pump, January 1, 1990; at C, acid concentrate was
initiated through the second pump, February 1, 1992; at D, use of both pumps was discontinued, February 7, 1992.

of the electric pumps, the epidemic period was defined as October Laboratory investigation of environmental samples
15, 1988 to February 7, 1992. To determine risk factors for
elevated aluminum levels during the epidemic period we ran- Samples of acetate and bicarbonate dialysate, and acid, bicar-
domly selected and compared 10 of the case- and 10 control- bonate and acetate concentrate were collected for determination
patients. Because one of the identified risk factors was not present of aluminum levels. Aluminum assays were performed at Damon
until after December 31, 1989, we next performed a cohort study Clinical Laboratories, Inc. (Smyrna, GA, USA) or Serco Labora-
of all DCA patients receiving dialysis during January 1, 1990 to tories (St. Paul, MN, USA) using the atomic absorption method
February 7, 1992, for a total of 102 (including 17 case-control [29, 30]. One DCA electric pump was sent to CDC for examina-
study) patients. Twelve patients who did not have an aluminum tion and testing; acid concentrate was passed through the pump
level after January 1, 1990 and four who had a level but not after and tested for aluminum.
5 dialysis sessions were excluded; most were only transiently
dialyzed at DCA (<2 months). Risk factors evaluated included
demographic factors, receipt of medications containing alumi- Laboratory investigation of autopsy specimens
num, and type of dialysate, dialyzer, and dialysis machine used.
Dialysis sessions included were from the beginning of the period Fresh autopsy specimens of brain from one patient suspected of
being evaluated (that is, October 15, 1988 for the case-control having aluminum intoxication were sent to National Medical
study or January 1, 1990 for the cohort study) until the date of the
Services, Inc. (Willow Grove, PA, USA) for aluminum testing.
patient's subsequent maximum aluminum level, before deferox-
amine.

Procedural investigation
We reviewed the design and maintenance of DCA dialysate 1 Use of company or trade names is for identification only and does not
delivery and water treatment systems and water quality monitor- imply endorsement by the Public Health Service or the U.S. Department
ing records. of Health and Human Services.
 Buiwen et al: Epidemic aluminum intoxication 471

Analysis of clinical characteristics of patients with Table 1. Comparison of potential risk factors among patients with a
serum aluminum level 100 jrg/liter' (case-patients) versus patients
aluminum exposure with maximum serum aluminum level <boa jig/liter (noncase-patients),
To determine the clinical characteristics of patients in the from Dialysis Center A, January 1, 1990—February 7, 1992
cohort study, we reviewed their outpatient and/or inpatient med- Continuous van ables [medians (25th, 75th percentiles)]
ical records. Data were collected on mental status change requir-
Potential risk Cases Noncases
ing hospitalization, seizures (generalized tonic-clonic convulsions factors (N = 59) (N = 27) P value
requiring hospitalization, partial seizures, or epileptiform activity
on electroencephalogram leading to anticonvulsant treatment), Number of dialysis 93 (44, 159) 34 (16, 118) 0.007
sessions
bone fractures, and receipt of blood transfusions, erythropoietin, Number of dialysis
iron dextran injections, or deferoxamine. Potential associations of sessions using
selected clinical characteristics with aluminum exposure (before COBE machine 6 (1, 20) 0 (0, 7) 0.0 12
onset) were evaluated. Based on our initial investigation, a dialysis SPS45O machine 69 (41, 129) 28 (6, 99) 0.010
session using acid concentrate pumped through the electric Bicarbonate 77 (35, 125) 0 (0, 18) <0.0001
pumps was defined as an aluminum exposure. Seizures or mental dialysate
Acid concentrate 45 (15, 84) 0 (0, 3) <0.0001
status change of clear etiology unrelated to aluminum (such as through
hypernatremia, sepsis) were not defined as a positive outcome in electric pump
the analysis. Acetate dialysate 0 (0, 9) 16 (0, 87) 0.013
Acetate 0 (0, 9) 16 (0, 87) 0.013
Patients with aluminum exposure were given deferoxamine concentrate
stimulation tests and treated when appropriate with intravenous through
0.5 to 1 g deferoxamine, usually administered weekly, until electric pump
stimulated serum aluminum levels were similar to unstimulated
levels. Patients present at DCA after February 7, 1992 were Categorical variables
followed for one year. Potential risk factors Rate RRb 95% CF P value
Predominant machine
Statistical methods SPS45O 52/73 (71%) 1.32 0.78—2.23 0.33
CUBE 7/13 (54%)
Data were entered into a microcomputer and analyzed using Predominant bath
Epi Info [311. The Fisher's exact or test was used for analysis of Bicarbonate dialysate 51/61 (84%) 2.61 1.46—4.68 <0.0001
categorical variables; the Kruskal-Wallis test was used for analysis Acetate dialysated 8/25 (32%)
of continuous variables. To assess the relationship between two Acid concentrate
variables, we calculated a Pearson's correlation coefficient by through electric
pump
linear regression analysis. Yes 52/62 (84%) 2.88 1.53—5.42 <0.0001
No 7/24 (29%)
Results Acetate concentrate
During the study period, 64 patients had a serum aluminum through electric
pump
level 100 jig/liter; many were identified in January 1992 due to Yes 22/40 (55%) 0.68 0.50—0.94 0.021
testing of all patients (Fig. 1). Thirteen of these had levels as high No 37/46 (80%)
as 300 to 500 jig/liter; six had levels >500 jig/liter. Serum aluminum level
The proportion of patients with serum aluminum levels 100 before 1/1/90
>100 jig/liter
jig/liter was higher during the period after the electric pumps were Yes 8/8 (100%) 1.53 1.30—1.80 0.052
installed (October 15, 1988 to March 26, 1992) than before they No 51/78 (65%)
were installed (January 1, 1987 to October 14, 1988); 60 of 133 a Before receipt of deferoxamine
(45%) versus 3 of 86 (3.5%), relative risk (RR) of 13, and P b Relative risk
<0.0001 (analysis excludes one patient with a first serum alumi- 95% confidence interval
d Acetate dialysis sessions bicarbonate dialysis sessions
num level 100 jig/liter before the study period). Fifty-nine
(92%) of the 64 patients with serum aluminum levels 100
jig/liter met the case definition.
higher number of sessions using SPS45O (Baxter Healthcare
Analysis of risk factors for elevated aluminum levels
Corporation, Largo, FL, USA) dialysis machines, receipt of
Case-control study. Case-patients were significantly more likely predominantly bicarbonate dialysate, and receipt of acid concen-
than controls to have a higher number of dialysis sessions using trate passed through an electric pump (Table 1). Age, gender,
bicarbonate dialysis (median 101 vs. 0, P = 0.01), higher number duration of dialysis at DCA, type of dialyzer, use of predominantly
of sessions using acid concentrate (used in bicarbonate dialysis) CUBE or SPS45O machines, or receipt of oral medications
passed through an electric pump (37.5 vs. 0, P = 0.02), or higher containing aluminum or citrate were not associated with becom-
number of sessions using COBE Centrysystem 2 (COBE Labora- ing a case-patient. Receipt of acetate concentrate through an
tories, Inc., Lakewood, CO, USA) dialysis machines (12.5 vs. 0, P electric pump, higher number of sessions using acetate dialysate,
= 0.01). and higher number of sessions using acetate concentrate through
Cohort study. Of 86 DCA patients included in the cohort study, an electric pump were protective factors.
59 met the case definition. Risk factors were all those identified in We controlled for the number of dialysis sessions by examining
the case-control study plus higher number of dialysis sessions, exposures only during the month before the maximum aluminum
472 Burwen et a!: Epidemic aluminum intoxication

Table 2. Dialysis fluid aluminum assays, Dialysis Center A On February 7, 1992, DCA discontinued use of the electric
Date of Aluminum pumps. On February 18, the storage tanks were emptied and
collection Specimen Source mg/liter cleaned; they were then refilled. On February 20, the electric
pumps were removed and replaced with plastic electric pumps
2/12/92 Acid concentrate storage tank 1a
2/12/92 Acid concentrate storage tank 1, and after 302 (model MDX-MT3, March Manufacturing, Inc., Glenview, IL,
passing through the USA). Each of the removed pumps had an aluminum casing,
aluminum pump casing cover, and impeller.
2/18/92 Acid concentrate storage tank 1, and after 2.9 The water treatment system included a reverse osmosis unit and
passing through the a back-up deionization unit. Routine testing of the treated water
aluminum pump
2/18/92 Acid concentrate storage tank 2b and after 0.05 1 (product water) for aluminum was performed once in 1987 and
passing through the approximately every six months during February 1989 to February
aluminum pump 1992; results always met AAMI standards [21, 29].
4/21/92 Acid concentrate storage tank 1, and after 0.007
passing through plastic
Laboratory investigation
pumpc
4/21/92 Acid concentrate storage tank 2, and after <0.005 Product water samples obtained in April 1992 during our
passing through plastic investigation had aluminum levels within AAMI standards. Alu-
pump minum levels in acid or acetate concentrate, bicarbonate solution,
4/21/92 Bicarbonate solution before entering machine 0.029
4/21/92 Bicarbonate dialysate before passing through <o.oos or dialysate were also low (Table 2). Acid concentrate samples
dialyzer collected in February 1992 after passing through the aluminum
4/21/92 City water 0.018 pump had variable aluminum levels depending upon the date of
4/21/92 Reverse osmosis <0.005
collection, but all levels were elevated. Based on the 302 and 2.9
product water
5/1/92 Acid concentrate CDC <0.005 mg/liter measurements, the calculated aluminum level in dialysate
5/1/92 Acid concentrate CDC; single pass through 200 after 1:36 dilution by the SPS45O dialysis machine is 81 to 8400
aluminum pump xgIliter. This value is 8.1 to 840 times the maximum AAMI
5/1/92 Acid concentrate CDC; after 3 days in 870 recommended aluminum level in product water. When the DCA
aluminum pump
pump was set up at CDC, acid concentrate passed once through
Used with SPS45O machines the pump, or left in the pump over a weekend and then flushed,
b Used with COBE machines
had high levels of aluminum. The pH of selected acid concentrate
Aluminum pumps removed 2/21/92
samples was 2.6 and for acetate concentrate was 6.7.

Laboratory investigation of autopsy specimens

level. Exposure to bicarbonate dialysis (55 of 67 vs. 4 of 19, RR =
Autopsy specimens of brain from one patient with encephalop-
3.9, P < 0.0001) or acid concentrate through an electric pump (48
of 56 vs. 11 of 30, RR = 2.3, P < 0.0001) were significant; acetate athy had an elevated level of aluminum in the gray matter
compared to that in the white matter; 8.2 and 2.0 pg/gm wet wt,
dialysate (3 of 20 vs. 56 of 66, RR = 0.18, P < 0.0001) and acetate
respectively. This patient had 172 dialysis sessions of aluminum
concentrate through an electric pump (3 of 20 vs. 56 of 66, RR =
0.18, P < 0.0001) were again protective factors. The type of exposure and a maximum serum aluminum level of 642 jLg/liter.
dialysis machine was not significant.
When we stratified by receipt of acid concentrate through an Analysis of clinical characteristics of patients with
electric pump, receipt of predominantly bicarbonate dialysis or aluminum exposure
use of predominantly COBE or SPS45O machines were not Patients in the cohort study had 0 to 181 (median 50) dialysis
significant. sessions with aluminum exposure. Seventeen (20%) had hospital-
The maximum serum aluminum level (before deferoxamine) i.zations for mental status change or seizures. These patients had
and number of dialysis sessions using acid concentrate through an significantly more dialysis sessions with aluminum exposure than
electric pump were positively correlated (r 0.63). did patients without these problems (mental status change: me-
dian 92.5 vs. 40.5, P 0.0006; seizure: 109.5 vs. 45.5, P = 0.002).
Environmental and procedural investigation Nine of the 17 patients had neurologic or systemic diseases
Before October 1988, acetate or acid concentrate was stored in unrelated to aluminum intoxication which could affect their
55 gallon drums and extracted using a hand pump. In October mental status, such as cerebrovascular disease, pre-existing sei-
1988, two 500-gallon storage tanks were introduced, which were zure disorder (previously controlled), amyloidosis, or pneumonia.
located at ground level. Two electric pumps (Teel Industrial Characteristics of the eight patients without identifiable pre-
Series model 1P862, Dayton Electric Manufacturing Company, existing neurologic problems included seizures in 5 (63%), en-
Chicago, IL, USA) were installed to pump fluid from the tanks to cephalopathy in 4 (50%), myoclonic jerks in 3 (38%), speech
the second floor of DCA, via polyvinylchloride and vinyl pipes. difficulty in 6 (75%), inappropriate behavior in 3 (38%), halluci-
From October 1988 through December 1989, both storage tanks nations in 2 (25%), dementia in 1 (13%), confusion in 7 (86%),
contained acetate concentrate. In January 1990, because an disorientation in 6 (75%), and lethargy in 4 (65%). Clinical
increasing number of patients were being treated with bicarbonate characteristics of the two groups did not differ significantly.
dialysate, one tank was converted to acid concentrate for use with Fifteen (88%) of the 17 patients with hospitalizations for mental
SPS45O dialysis machines. On February 1, 1992, the second tank status change or seizures had onset before or within a few days
was coaverted to acid concentrate, for use with COBE machines. after February 7, 1992, when use of the aluminum pumps was
 Burwen et al: Epidemic aluminum intoxication 473

discontinued. The other two patients had onset two months after accounted for their subsequent abnormalities or predisposed
this date. Bone fractures, or receipt of blood transfusions, eryth- them to clinical manifestations of aluminum intoxication.
ropoietin or iron dextran injection were not associated with Persons involved in the operation of dialysis centers should be
aluminum exposure. aware of the potential for contamination of dialysate or water by
aluminum or other chemicals. Flouride, chloramine, sulfate,
nitrate, zinc, sodium azide, and copper can cause toxicity in
Discussion dialysis patients, and many other chemicals pose a potential
hazard [21, 29, 37—39]. Materials used in water treatment and
Epidemic aluminum intoxication occurred in DCA in associa- dialysate delivery systems, including fluid pathways, storage tanks,
tion with the increased use of bicarbonate dialysis and the delivery central delivery systems, and pumps, should be compatible with all
of acid concentrate through an aluminum pump. Unlike previ- fluids in the system [21, 29, 37]. Modifications of the water
ously reported epidemics, this outbreak occurred while an effec- treatment, dialysate delivery, or dialyzer reprocessing systems
tive water treatment system was in place, and was due instead to should only be made by personnel knowledgable about these
a device used in the dialysate delivery system. The potential for systems and AAMI standards for water and dialysate used in
similar epidemics may be increased because the percentage of hemodialysis. Based on this outbreak, the Food and Drug Admin-
U.S. dialysis centers using bicarbonate-based dialysate as their istration sent a Safety Alert to U.S. dialysis centers recommending
primary method of dialysis has increased from 22% in 1986 to reassessment of the design of and components used in the water
92% in 1991 [32]. treatment, concentrate delivery transfer/storage, and dialysate
The two aluminum pumps installed at DCA in October 1988 delivery systems whenever changing or updating any component
were initially used to deliver acetate concentrate, which has a of an existing dialysis unit, and routine monitoring of serum
neutral pH. Subsequently, acid concentrate for 5P5450 machines aluminum levels in dialysis patients [40].
was delivered through one aluminum pump during January 1990
to February 1992. Acid concentrate was used with the second Acknowledgments
aluminum pump only during a one week period in February 1992. We thank the unit administrator, medical director, and other personnel
Evidence implicating the aluminum pumps, and specifically the at DCA for their help in the conduct of this outbreak investigation; the
delivery of acid concentrate through one pump over a two year neurologist who performed the follow-up patient evaluations; Jerome
period includes the strong epidemiologic associations with receipt Tokars, M.D., M.P.H. (CDC) for technical assistance, and Linda WaIler
of acid concentrate through a pump, receipt of bicarbonate (CDC) for graphic assistance.
dialysate, higher number of sessions using 5P5450 machines, and
Reprint requests to William R. Jarvis, MD., Hospital Infections Program,
the correlation between serum aluminum level and higher number Centers for Disease Control and Prevention, Mailstop E69, 1600 Clifton Rd.
of sessions with acid concentrate through a pump which suggests NE, Atlanta, Georgia 30333, USA.
a dose-response. Laboratory results showing high levels of alumi-
num in acid concentrate which had passed through one pump and References
low levels in DCA samples taken after the pump was removed
confirmed the aluminum-containing pump as the source of chem- 1. ALPREY AC: Dialysis encephalopathy. Kidney mt 29(Suppl 18):S53—
557, 1986
ical contamination. 2. SHERRARD DJ, ANDRESS DL: Aluminum-related osteodystrophy. Adv
Factors contributing to the perpetuation of the outbreak over a Intern Med 34:307—324, 1989
two year period included case-patients being distributed among 3. Bit.. MJ, COOPER K, SCHNALL 5, Dunn' T, HENDLER E, MALLUCHE H,
physician groups so individual physicians were not aware of the SOLOMON L: Aluminum induced anemia: Pathogenesis and treatment
in patients on chronic hemodialysis. Kidney mt 36:852—858, 1989
extent of the outbreak, variability in patients' serum aluminum 4. SCHORR W: Disturbances of calcium metabolism in patients on
levels causing confusion about their clinical significance, and maintenance haemodialysis. Proc Eur Dial Transplant Assoc 5:408—
receipt of oral aluminum phosphate binders by some patients who 411, 1968
seemed therefore to have a plausible cause for their elevated 5. KERR DNS, WAaS J, ELLIS H, SIMPsoN W, ULDALL PR, WAiw MK:
serum aluminum level. Discontinuation of aluminum containing Bone disease in patients undergoing regular haemodialysis. (abstract)
JBone Joint Surg 51B:578, 1969
medications was often done concomitantly with a change in 6. ALPREY AC, MISHELL JM, BIRKS J, CONTIGUGLIA SR, RUDOLPH H,
patients' dialysate to a low calcium bath. Aluminum levels im- LEwIN E, HOLMES JH: Syndrome of dyspraxia and multifocal seizures
proved, reinforcing the belief that discontinuation of these med- associated with chronic hemodialysis. Trans Am Soc Amf Intern Org
ications controlled the problem; however, in retrospect, the 18:257—261, 1972
7. SIMPSON W, KERR DNS, HILL AVL, SIDDIOUI JV: Skeletal changes in
improvement was probably mainly due to the new bath which did
patients on regular hemodialysis. Radiology 107:313—320, 1973
not use acid concentrate through the aluminum pump. 8. BURKS JS, ALPREY AC, HUDDLESTONE J, NORENBERG MD, LEwIN E:
Results of aluminum testing of autopsy specimens of brain A fatal encephalopathy in chronic haemodialysis patients. Lancet
tissue available from one patient were consistent with aluminum i:764—768, 1976
encephalopathy [33—36]. Although similar data are not available 9. Pi.xnS MM, GOODE GC, HISLOP JS: Composition ot the domestic
water supply and the incidence of fractures and encephalopathy in
for other patients, the epidemiologic association of seizures or patients on home dialysis. Br Med J 11:657—660, 1977
mental status change with aluminum exposure supports our belief 10. ELuon HL, DRYBURGH F, FELL GS, St..nET 5, MACDOUGALL AL
that the neurologic problems of a number of other patients were Aluminium toxicity during regular haemodialysis. Br Med J 1:1101—
due to aluminum intoxication. In eight patients, these manifesta- 1103, 1978
11. DUNEA G, MAHURKAR SD, MAMDANI B, SMITH EC: Role of alumi-
tions had no clear etiology and in retrospect may have been due num in dialysis dementia. Ann mt Med 88:502—504, 1978
to aluminum intoxication. Another nine patients had concomitant 12. RozM VV, PORT FK, Run WM: Progressive dialysis encephalopathy
neurologic or systemic diseases that may have either solely from dialysate aluminum. Arch Intern Med 138:1375—1377, 1978
474 Bunven et al: Epidemic aluminum intoxication
13. RozM VV, PORT FK, EASTERLING RE: An outbreak of dialysis
dementia due to aluminum in the dialysate. J Dial 2:459—470, 1978
14. PARKINsON IS, WARD MK, FEasT TO, FAWCETr RWP, KERR DNS:
Fracturing dialysis osteodystrophy and dialysis encephalopathy: an
epidemiological survey. Lancet i:406—409, 1979
15. PmRIDE5 AM, EDWARDs WG JR, CULLUM UX JR, MCCALL JT, Ews
HA: Hemodialysis encephalopathy with osteomalacic fractures and
muscle weakness. Kidney mt 18:115—124, 1980
16. Report from the Registration Committee of the European Dialysis
and Transplant Association. Dialysis Dementia in Europe. Lancet
ii:t90—192, 1980
17. LEATHER HM, LEwIN 10, CALDER E, BRAYEROOKE I, Cox RR: Effect
of water dcionisers on 'fracturing osteodystrophy' and dialysis enceph-
alopathy in Plymouth. Nephron 29:80—84, 1981
18. DAvisoN AM, Ou H, WALKER OS, LEwIN5 AM: Water supply
aluminum concentration, dialysis dementia, and effect of reverse-
osmosis water treatment. Lancet ii:785—787, 1982
19. SCHREEDER MT, FAVERO MS, HUGHES JR. PETERSEN NJ, BENNErF
PH, MAYNARD JE: Dialysis encephalopathy and aluminum exposure:
An epidemiologic analysis. J Chron Dis 36:581—593, 1983
20. MILNE FJ, SFIARF B, BELL P, MEYERS AM: The effect of low aluminum
water and desferrioxaminc on the outcome of dialysis encephalopathy.
Clin Nephrol 20:202—207, 1983
21. A5S0cL4TI0N FOR THE ADVANCEMENT OF MEDICAL INSTRUMENTA-
TION: American National Standard for Hemodialysis Systems (ANSI!
AAMI RD5—1 981). Arlington, 1981
22. FLENDRIG JA, KRUIS H, DAS HA: Aluminium intoxication: The cause
of dialysis dementia? Proc Eur Dial TransplantAssoc 13:355—368, 1976
23. PIERIDES AM, FROI-INERT PP, SHARBROUGH FW: Release of alumi-
num by the Redy dialysis cartridge: Osteomalacia and dialysis enceph-
alopathy after prolonged usage. Blood Punf 1:145—153, 1983
24. NORRIS KC, CROOKS PW, NEBEKER HG, HERCZ 0, MILLINER DS,
GERSZI K, SLATOPOLSKY E, ANDRESS DL, SHERRARD DJ, COBURN
JW: Clinical and laboratory features of aluminum-related bone dis-
ease: Difference between sporadic and "epidemic" forms of the
disease, Am J Kidney Dis 6:342—347, 1985
25. ANDREOLI SP, BERGSTEIN JM, SI-IER1WW DJ: Aluminum intoxication
from aluminum-containing phosphate binders in children with
azotemia not undergoing hemodialysis. N Engl I Med 310:1079—1084,
1984
26. BAKIR AA, HRYHORCZUK DO, BERMAN E, DUNEA 0: Acute fatal
hyperaluminemic encephalop'tthy in undialyzed and recently dialyzed
uremic patients. ASAIO Trans 32:171—176, 1986
27. ROEERTSON JA, SALUSKY lB. GOODMAN WO, NORRIS ICC, COBURN
JW: Sucralfate, intestinal aluminum absorption, and aluminum toxic-
ity in a patient on dialysis. Ann Intern Med 111:179—181, 1989
28. BURGESS E: Aluminum toxicity from oral sucralfate therapy. Nephron
59:523—524, 1991
29. ASSOQArI0N FOR THE ADVANCEMENT OF MEDICAL INSTRUMENTA-
TION: American national standard for hemodialysis systems (ANSI!
AAMI RDS—1992). Arlington, 1992
30. U. S. ENVIRONMENTAL PROTECTION AGENCY: Methods for Chemical
Analysis of Water and Waste (EPA-600—4-79—020,). Cincinnati, Envi-
ronmental Research Center, Environmental Monitoring and Support
Laboratory, 1979 (revised March 1983)
31. DEAN AD, DEAN JA, BURTON JH, DICKER RC: Epi Info, Version
5.Olb: A Word Processin& Database, and Statistics Program for Epide-
miology on Microcomputers. Atlanta, Centers for Disease Control,
1990
32. TOKARS JI, ALTER MJ, FAVERO MS, MOYER LA, BLAND LA: National
surveillance of dialysis associated diseases in the United States, 1991.
ASAIO 1 39:966—975, 1993
33. ALFREY AC, LEGENDIth OR, KAEHNY WD: The dialysis encephalop-
athy syndrome: Possible aluminum intoxication. N Engl I Med 294:
184—188, 1976
34. MCDERMOn JR, SMITH Al, Wium MK, PARKINSON IS, KERR DNS:
Brain-aluminum concentration in dialysis encephalopathy. Lancet
1:901—903, 1978
35. CARTIER F, ALLAIN P, GARY J, CHATEL M, MENAULT F, PECKER 5:
Encbphalopathie myoclonique progressive des dialyses. Role de l'eau
utilisée pour l'hCmodialyse. Nouv Presse Med 7:97—102, 1978
36. ALFREY AC: Aluminum metabolism in uremia. Neurotoxicology 1:43—
53, 1980
37. ARDUINO MJ, BLAND LA, FAVERO MS: Adverse patient reactions due
to chemical contamination of hemodialysis fluids. Dial Transplant
18:655—658, 1989
38. FAVERO MS, ALTER MJ, BLAND LA: Dialysis-associated infections
and their control, in Hospital Infections (3rd ed), edited by JV
BENNETr, P BRACHMAN, Boston, Little, Brown and Co., 1992, pp
375—403
39. ARNOw PM, BLAND LA, GARCIA-HOUCHINS 5, FRIDKIN 5, FELLNER
SK: An outbreak of fatal fluoride intoxication in a long-term hemo-
dialysis unit. Ann Intern Med 121:339—344, 1994
40. FDA Safety Alert: Aluminum and Other Trace Element Contamination
in Dialysis Facilities. U.S. Department of Health and Human Services,
Public Health Service, Food and Drug Administration, May 20, 1992