AIDS Dementia Complex
by Mark Bowers
HIV is known to affect most if not all systems in the body. A member
of the subfamily of retroviruses that causes a variety of neurological
and immunological diseases, HIV has been found in the brain as early as
2 days after initial infection. A type of dementia usually referred to
as AIDS dementia complex (ADC), but also known as HIV-1-associated dementia
or HIV-associated cognitive/motor complex, has been estimated to affect
up to one-third of adults and one-half of all children with AIDS.
It remains unclear exactly how and where HIV enters the brain. However,
newly devised methods of measuring HIV in the cerebrospinal fluid (CSF)
that bathes the spinal cord and the brain may be predictive of the risk
of developing ADC and may help gauge efforts to prevent or treat it. The
availability of antiretroviral drugs from 2 new classes, protease inhibitors
and non-nucleoside reverse transcriptase inhibitors (NNRTI), potentially
increases the number of options for preventing and treating ADC. New and
better understanding of how HIV causes disease in the brain opens the
door for testing novel therapies, including anti-inflammatory drugs, antioxidants,
cytokine regulators and calcium channel blockers.
It has been reported that reductions in viral load as measured by polymerase
chain reaction (PCR) and branched-chain DNA (bDNA) assays have led to
decreases in the incidence of new opportunistic infections (OI) and in
the total number of AIDS-related hospitalizations among individuals who
use powerful combinations of antiretroviral drugs. Will there also be
decreases in the number of people who develop dementia and related cognitive
deficits, or will clinicians see more dementia because people with HIV
are living longer? Is the brain a reservoir site where HIV persists, even
while it is reduced to undetectable levels in the blood? What can be done
now to prevent and treat ADC, and what remains to be done in the near
future?
Dementia and its Symptoms
The term dementia is used by neurologists to describe a clinical syndrome
composed of memory loss, decreased mental concentration and the loss of
other intellectual functions because of progressive disease in the brain.
There may be motor signs, such as weakness, lack of coordination and unsteady
gait. Behavioral changes are most likely to be noticed by friends or family
members, and may include apathy, personality changes and loss of libido.
ADC is often characterized by depression, but depression alone is not
enough to establish a diagnosis of ADC. Depression has a profound effect
on survival and should be treated, whether it appears as part of the constellation
of symptoms called ADC or alone.
The clinical course of ADC is variable. Some people experience only mild
symptoms, while others progress rapidly. A scale that allows clinicians
to stage ADC is provided below. The first symptoms to appear are often
the most difficult to quantify: short-term memory loss and poor concentration
may be attributed to stress, to prescription or recreational drug use,
or to fatigue, but may also be early manifestations of ADC.
AIDS
Dementia Complex Classification
- Stage 0: Normal
- Stage 0.5: Subclinical or Equivocal
- Minimal or equivocal symptoms
- Mild (soft) neurological signs
- No impairment of work or activities of daily living (ADL)
- Stage 1: Mild
- Unequivocal intellectual or motor impairment
- Able to do all but the most demanding work or ADL
- Stage 2: Moderate
- Cannot work or perform demanding ADL
- Capable of self-care
- Ambulatory, but may need a single prop
- Stage 3: Severe
- Major intellectual disability or
- Cannot walk unassisted
- Stage 4: End-Stage
- Nearly vegetative
- Rudimentary cognition
- Para- or quadriplegic
Adapted from J Worley and R Price, Management of neurologic
complications of HIV-1 infection and AIDS. In The Medical Management of
AIDS, 3rd Edition. M Sande and P Volberding, editors. WB Saunders Company,
Philadelphia, 1992.
Many AIDS-related OI, including toxoplasmosis, cryptococcal meningitis,
cytomegalovirus (CMV) encephalitis and progressive multifocal leukoencephalopathy
(PML) have symptoms similar to those caused by ADC. An accurate diagnosis
based on neurological, cognitive and psychological evaluations can help
sort out the underlying causes. Other scales have been proposed for staging
ADC than the one presented above; the scale developed for minor cognitive/motor
disorder is designed to capture and stage persons with early cognitive
difficulties, and overlaps the first stage of the ADC scale. The point
of creating the new scale was to draw attention to cognitive difficulties
and to attempt to intervene early in the course of ADC.
HIV infection of the central nervous system (CNS) in children differs
from the picture seen in adults. Children frequently fail to reach normal
developmental milestones or achieve normal neurologic development. The
effects of CNS involvement are seen early. However, among adults, AIDS
dementia complex is principally a late manifestation of HIV infection,
most frequently occurring in individuals with fewer than 200 CD4 cells/mm3.
Nevertheless, early symptoms of ADC may arise at any time during HIV infection,
and should be addressed when they arise. There are promising new treatments
on the horizon, and the possibility of halting or reversing the progression
of dementia is very real.
Diagnosis
A diagnosis of AIDS dementia complex is based on excluding other common
causes of the symptoms seen. ADC is referred to as a subcortical dementia
because it affects structures that lie under the surface of the brain,
the cortex. In persons who have ADC, increased metabolism has been recorded
from subcortical parts of the brain called the thalamus and the basal
ganglia. The basal ganglia support muscular activities, including posture,
balance and walking. Neurologists look for both direct and indirect evidence
of dementia by a variety of diagnostic measures, including neurological,
cognitive and psychological tests.
Lumbar punctures (spinal taps), in which a needle is inserted into the
CSF surrounding the spinal cord, allow the withdrawal of a small amount
of the fluid to test for the presence of active infections in addition
to HIV, including Cryptococcus neoformans, Treponema pallidum
(the organism that causes syphillis), CMV, Epstein-Barr virus, JC virus
and Mycobacterium tuberculosis. CSF removed by lumbar punctures
may also provide valuable information about levels of HIV RNA and DNA.
Researchers want to know if these levels change in response to antiviral
therapy, and if the changes reflect successful therapy for ADC. Lumbar
punctures are quite safe. The primary side effect is headache if the patient
does not remain lying down long enough after the procedure.
Researchers have collected data about the amounts and distribution of
beta-2 microglobulin and neopterin in CSF and the brain. Bruce Brew, MD,
and colleagues at St. Vincent's Hospital in Sydney, Australia, found a
correlation between concentrations of beta-2 microglobulin and neopterin
in CSF, CD4 cell counts and the development of ADC. A total of 35 participants
with ADC stage 0.5 or without ADC and CD4 cell counts less than 200 cells/mm3
were evaluated in the study. Brew found that elevated concentrations of
CSF beta-2 microglobulin and neopterin "are significant predictors
of the risk of development of ADC."
Single photon emission computed tomographic (SPECT) cerebral perfusion
scans and magnetic resonance imaging (MRI) brain scans are useful tools
for establishing ADC and ruling out other possible brain complications,
such as CMV-associated CNS disease, toxoplasmosis, or subcortical lesions
or processes. Abnormal glucose metabolism in patterns consistent with
early or late ADC can be detected in most affected individuals by positron
emission tomography (PET) scanning. Computerized tomography (CT scan)
is useful for diagnosing toxoplasmosis, but is not helpful in establishing
a diagnosis of ADC.
David Rottenberg, MD, and colleagues at the University of Minnesota used
FDG-PET scans and compared blood samples from both HIV positive and HIV
negative volunteers to expand clinical knowledge about hypermetabolism
(high metabolism) of glucose in the areas where cellular abnormalities
first appear in people with ADC and hypometabolism (low metabolism) in
the cortex of people with more advanced ADC. The correlations were robust;
now further testing of this experimental technology is needed to establish
whether FDG-PET scans can predict progression of ADC or response to therapy.
How Does HIV Affect the Brain?
Neurologists are convinced that HIV infection in the brain is indirectly
responsible for AIDS dementia complex. Accumulating evidence shows that
many types of cells in the brain can be and are infected by HIV, but neurons
themselves are not. Although HIV does not directly infect and kill neurons,
neuronal death has been repeatedly associated with ADC. How then does
HIV cause ADC?
Indirect killing of neurons by a toxic substance or substances released
by HIV-infected cells is the likely explanation for the emergence of ADC.
The cells that are infected in the CNS are microglia (relatives of macrophages),
vascular endothelial cells that line the blood vessels of the brain, and
astrocytes, cells that fill the spaces between neurons in the brain and
provide support and insulation for neurons.
HIV infection of brain cells may cause collateral damage to neighboring
neurons through the release of toxic chemical messengers called cytokines.
Some of the cytokines that are produced abundantly by infected cells include
tumor necrosis factor alpha (TNF alpha), gamma and alpha interferon (IFN
gamma and IFN alpha) and platelet activating factor (PAF). Nitric oxide
(NO), a reactive nitrogen intermediate (free radical) that is reponsible
for vasodilation (dilation of blood vessels), and metabolites such as
arachidonic acid are also abundantly produced during gp120-induced inflammation.
A general pattern of dysregulation accompanies clinical dementia.
The hypothesis that such substances lead to ADC can be tested by giving
drugs that decrease levels of substances such as TNF alpha or PAF to patients
with ADC and seeing whether they improve. Some drugs that are in clinical
study for this purpose are the PAF inhibitor lexipatant (made by British
Biotech for the treatment of pancreatitis, asthma and multiple sclerosis),
and pentoxifylline (Trental), a potent TNF alpha inhibitor. A putative
antioxidant, OPC 14,117, was studied by Karl Kieburtz, of the University
of Rochester, and colleagues in 30 people with fewer than 245 CD4 cells/mm3
and some cognitive impairment. Results were not statistically significant,
but there was a trend toward improvement in memory and motor speed for
those who took OPC 14,117.
Neuronal damage may also be caused by the inflammation that results from
overproduction of cytokines. Astrocytes that produce TNF alpha enter a
feedback loop that causes them to continue to produce more and more astrocytes.
The resulting astrogliosis (proliferation of astocytes) may be in part
responsible for ADC, and contributes to neuron death and demyelination
(loss of important insulation surrounding neurons). Astrocytes communicate
with one another through what are called gap junctions, so damage that
occurs as the result of the infection of one astrocyte can lead to damage
at regions distant from the site of exposure to HIV.
Neuronal suicide (apoptosis, or programmed cell death) may be caused
by free-floating gp120, part of the external envelope of HIV, in the CNS.
Infected macrophages, monocytes and microglia all release gp120. An abundance
of gp120 in the CNS triggers a disruption in the normal concentrations
of extracellular substances such as glutamate and activates neural receptor
channels called NMDA channels. Neurons have specialized channels that
control the entry and exit of calcium (Ca), potassium (K) and sodium (Na)
ions, which allow these cells to communicate with other cells through
electrical polarization and depolarization. Opening a Ca channel for too
long is fatal to a neuron.
The interactions between cells of the CNS create a finely balanced system
that permits emotion, movement and thought. The system can be imbalanced
by HIV infection of astrocytes, macrophages/microglia and T-cells in the
CNS. Eventually, the changing microenvironment results in neurotoxicity
that produces the cognitive, motor and behavioral changes associated with
AIDS dementia complex.
Viral Load
Until recently, most researchers believed that HIV disease was characterized
by a long period of clinical latency during which viral production was
low. The development of sensitive viral load testing has allowed researchers
to accurately quantify virus in the blood. How do HIV RNA levels in the
CSF compare with plasma levels? Are elevated levels of HIV in the CSF
or brain responsible for ADC?
Richard Price, MD, Chief of the Neurology Service at San Francisco General
Hospital (SFGH), and Silvija Staprans, PhD, Director of the Virology Laboratory
at SFGH, want to learn the effects of protease inhibitors and NNRTI on
viral load in the CSF and on the course of ADC. According to Price, the
leading neuropathological hypothesis relates ADC to HIV infection in the
brain. Aggressive use of antiretroviral drugs could help alleviate dementia
in 2 ways: (1) the drugs get into the CNS to treat HIV infection, and/or
(2) systemic viral load reduction and reconstitution of body-wide immunity
have a positive influence on the brain as well.
The endothelial cells that line the capillaries of the brain are tightly
linked to form a blood-brain barrier, which is selectively permeable to
substances in the blood. Current scant data seem to indicate that protease
inhibitors do not cross the blood-brain barrier. However, anecdotal reports
suggest that those who do well systemically (those who achieve a pronounced
decrease in viral burden in the blood) tend not to develop ADC, which
supports hypothesis 2.
Price and Staprans first want to know if high levels of HIV in the CSF
correspond with increased incidence of ADC. (Viral load in the CSF and
in the brain may differ.) Price and Staprans feel that they are now in
a position to learn much about the natural history of ADC, about prevention
and treatment and about the rate at which resistance to protease inhibitors
develops in the brain and CSF. They recently received University of California
approval to conduct a series of lumbar punctures in individuals who are
initiating combination antiretroviral therapy. Among the questions to
be answered are:
- how quickly does the level of virus in CSF respond to systemic therapy
with protease inhibitors?
- does viral load in CSF change proportionally with viral load in peripheral
blood?
- do CNS-penetrating drugs have a more profound effect than nonpenetrating
drugs on HIV replication in the CSF?
Bruce Brew has looked at the CSF for evidence that higher levels of HIV
RNA correlate with severity of ADC. He and his associates evaluated CSF
from 26 volunteers, 10 of whom were diagnosed with stage 1 or stage 2
ADC and found good correlations. However, elevated levels of HIV RNA could
also be associated with cryptococcal meningitis. (Possibly this increased
HIV replication is due to immune activation because of a simultaneous
infection.) BrewÕs findings suggest a place for the measurement of HIV
RNA in diagnosing dementia, but a single measure of elevated viral load
is not enough to distinguish ADC from other brain abnormalities.
Treatment and Prevention
Three current approaches to the treatment of ADC are:
- to interfere with the production of HIV through the use of antiviral
drug combinations in the expectation that systemic decreases in HIV
viral load will lead to a decrease in the incidence of ADC;
- to protect neurons from collateral damage by blocking the effects
of dysregulated cytokines; and
- to offset the neuronal damage caused by elevated concentrations of
nitric oxide in the brain with anti-inflammatory and antioxidant drugs.
Although there is no approved therapy for ADC, AZT is the drug that has
been most studied for its effects on ADC. AZT crosses the blood-brain
barrier better than any other approved nucleoside analog drug. Richard
Price points out that the evidence for an AZT effect on the development
and course of ADC is strongest in children. There is some evidence that
ADC-effective doses of AZT are much higher than those usually prescribed:
1,000-2,000 mg per day. According to Seth Hetherington, MD, a senior clinical
researcher and pediatric infectious disease specialist at Glaxo Wellcome,
a dose-response relationship for AZT is not well established. A study
of 40 volunteers with diagnosed ADC (stages 1 or 2) showed definite improvements
in neuropsychological testing scores after taking as much as 400 mg of
AZT 5 times a day. However, a 30-person, open-label Italian study comparing
1,000 mg, 750 mg and 500 mg per day showed cognitive improvement in all
participants taking AZT regardless of dose. Some progressed from stage
3 dementia to stage 1 or stage 0.5. Direct introduction of AZT into the
CSF was studied in 8 people, of whom 5 showed definite improvement in
ADC symptoms.
AIDS Clinical Trials Group (ACTG) 152, a study of AZT vs ddI vs the combination
of AZT plus ddI in symptomatic HIV-infected children, found that the second
most commonly reported clinical finding among children was neurologic
deterioration. Hetherington argues that it is necessary to develop drugs
with protective or therapeutic effects on the CNS with children as well
as adults in mind.
CNS
Penetration of Nucleoside Analog Drugs
(percentage of level found in serum)
AZT (Retrovir): 60%
ddI (Videx): 20%
ddC (Hivid): 20%
d4T (Zerit): 30-40%
3TC (Epivir): 10%
Other, newer antiretroviral drugs potentially could affect the course
of ADC, either as prophylactics or as treatments. William Freimuth, MD,
of Upjohn Pharmaceuticals, has shepherded the experimental NNRTI drug
delavirdine through the development and approval process. Freimuth comments
that early animal studies of delavirdine revealed that the concentration
of the drug in brain tissue is 5-10 times greater than its concentration
in the CSF. Penetration of delavirdine into the CSF is 0.4% of penetration
in plasma, which is not impressively high. However, a pilot study of delavirdine
for the treatment of ADC is ongoing, and an interim analysis is near.
The related NNRTI drug atevirdine was studied in 10 patients with stage
1 or 2 ADC. Five participants finished the 12-week study, 4 of whom responded
to atevirdine as measured by neurological and neuropsychological assessments.
Two participants improved from stage 2 to stage 0, while the other 2 improved
from stage 1 to stage 0.5. The recently approved NNRTI drug nevirapine
(Viramune, made by Boehringer Ingelheim) was evaluated early for its ability
to cross the blood-brain barrier. As a class, NNRTI drugs hold some promise
for the prevention and treatment of ADC, and should be evaluated appropriately.
Glaxo Wellcome is developing a new nucleoside analog drug, designated
1592U89 (or 1592), for the treatment of HIV disease. In Phase II testing
it was determined that levels of 1592 in the CSF reach 20% of their levels
in the blood an hour after an oral dose, then reach the same level of
penetration as AZT. Phase III testing of 1592 includes plans to evaluate
the effects of 1592 on 90 people with early signs of ADC. Study sites
include the University of California in San Francisco, the University
of California in San Diego, Johns Hopkins in Baltimore, Mount Sinai in
New York and Washington University in St. Louis. In parallel with adult
testing, pediatric trials will look at the effects of 1592 on neural development
in children with HIV.
The new protease inhibitor in joint development by Vertex and Glaxo Wellcome,
designated 141, showed initial high levels of penetration in the CNS of
rats. As reported at the Interscience Conference on Antimicrobial Agents
and Chemotherapy in New Orleans (September 15-18), penetration into the
CSF of this drug in humans was only 1.3% of blood levels, which is not
as impressive as the penetration of some NNRTI drugs. With all protease
inhibitors, penetration to the brain is minimal.
Drugs that inhibit neuronal apoptosis may be potential therapies for
ADC. Memantine has been shown to block the binding of gp120 to NMDA receptors;
this drug will be studied for its effects on ADC by the ACTG. A study
of the safety of memantine in people with HIV, lasting 13 weeks and offering
$1,025 for completion of 32 visits, is being conducted by ViRx in San
Francisco.
Another candidate for study is the familiar drug nitroglycerin, which
potentially could protect neurons from the effects of overstimulation
of NMDA receptors. The calcium channel blocker nimodipine (Nimotop, manufactured
by Miles Pharmaceuticals) was found in recent clinical studies to be unsatisfactory
in treating ADC. Another calcium channel blocker, verapamil, increases
HIV replication in vitro, and therefore appears unsuitable for clinical
study.
ACTG researchers are considering the anti-inflammatory corticosteroid
hormone prednisone for treating ADC. However, cytokine blockers and anti-inflammatory
drugs are expected to affect symptoms rather than the underlying cause
of ADC.
Peptide T, a peptide made of 8 amino acids, has been shown to block gp120
binding to target neural cells in some clinical studies. Some reports
have indicated that peptide T has led to improvements in neuropsychological
performance in people with severe ADC. However, researchers at the Curatorium
for Immunodeficiency in Munich, Germany, conducted a double-blind, randomized,
placebo-controlled study of peptide T in 128 volunteers, reported at the
XI International Conference on AIDS in Vancouver in July. Volunteers received
subcutaneous peptide T at 8.5 mg/day or placebo, and completed quality
of life questionnaires. Contrary to earlier reports of a positive effect,
there were no statistically significant effects of peptide T on quality
of life, neuropsychological performance or painful neuropathy.
Current and Future Clinical Research
The direction of future research will be determined by how well researchers
are able to find volunteers who have ADC. For the first time, researchers
and HIV-infected individuals have access to powerful antiretroviral drug
combinations that may affect the development of ADC (and other opportunistic
infections). Will these drugs be powerful enough to prevent dementia in
most HIV positive individuals, or will other interventions be necessary?
Or will HIV disease be marked by continual cognitive and intellectual
impairment as HIV positive people live longer, otherwise healthier lives?
The future of federally funded research on ADC is uncertain. Traditionally
neglected by the HIV/AIDS research community, neurologic research was
relegated to rare substudies or no studies at all by the ACTG, and received
scant attention from the National Institute of Neurologic Disease and
Stroke (NINDS). The Neurology AIDS Research Consortium (NARC) was created
2 years ago as a cooperative venture between NINDS and ACTG, and currently
sponsors and conducts most research on ADC. Due for a funding review in
late 1996, NARC must demonstrate the importance of including neurologic
research in future plans for evaluating combinations of protease inhibitors,
NNRTI and new nucleoside analog drugs.
Pharmaceutical companies play an important role as well. In the expectation
that HIV resistance to currently approved drugs will develop rapidly,
pharmaceutical companies continue pre-clinical evaluation of new drugs
with new mechanisms. They should include in their criteria for development
those drugs that can cross the blood-brain barrier. Freimuth has suggested
that Upjohn will evaluate future drugs for this ability. Lynn Smiley,
MD, International Director of Antiretroviral Clinical Research at Glaxo
Wellcome, points out that the new experimental drug 1592 was selected
for development precisely because of its ability to cross the blood-brain
barrier and enter the brain, a potential sequestered sanctuary site for
HIV.
Silvija Staprans suggests that primate research might provide clues about
the natural history of simian immunodeficiency virus (SIV) infection in
the brain, and could answer questions about the effects of drugs on SIV
and HIV replication. According to Staprans, "It is critical that
we understand the longevity of HIV-infected cells in the central nervous
system in order to better design rational therapeutic interventions for
durable suppression or eradication of HIV. And it is important to understand
whether the central nervous system is a focus for the development of drug-resistant
virus."
Mark Bowers is Managing Editor of Treatment Publications at the San
Francisco AIDS Foundation.
The author is indebted to Silvija Staprans, PhD, for independent editorial
review of the content of this article.
References
Brew B and others. Pilot study of the efficacy of atevirdine
in the treatment of AIDS dementia complex. AIDS 10:1357-1360. 1996.
Brew B and others. Predictive markers of AIDS dementia
complex: CD4 cell count and cerebrospinal fluid concentrations of beta-2
microglobulin and neopterin. Journal of Infectious Diseases 174:
294-8. 1996.
Cinque P and others. Polymerase chain reaction on cerebrospinal
fluid for diagnosis of virus-associated opportunistic diseases of the
central nervous system in HIV-infected patients. AIDS 10:951-958.
1996.
Epstein L and Gendelman H. Human immunodeficiency virus
type 1 infection of the nervous system: pathogenic mechanisms. Annals
of Neurology 33(5):429-436. May 1993.
Garth V and others. Efficacy of peptide T in the palliative
treatment of neurological HIV-related symptoms. XI International Conference
on AIDS. Vancouver, BC. July 1996. Abstract #Th.B.180.
Geldmacher D and others. Evaluation of dementia. New
England Journal of Medicine 335(5):330-336. August 1, 1996.
Johnson R. Quantitation of human immunodeficiency virus
in brains of demented and nondemented patients with acquired immunodeficiency
syndrome. Annals of Neurology 39: 392-395. March 1996.
Lipton S and Gendelman H. Dementia associated with the
acquired immunodeficiency syndrome. New England Journal of Medicine
332 (14):934-940. 6 April 1995.
Mascolini M. HIV and the Mind. Journal of the International
Association of Physicians in AIDS Care. June 1996.
Mascolini M. An interview with Justin C. McArthur, MB,
BS, MPH. Journal of the International Association of Physicians in
AIDS Care. June 1996.
Perry S. Organic mental disorders caused by HIV: update
on early diagnosis and treatment. American Journal of Psychiatry
147(6):696-710. June 1990.
Power C and others. HIV dementia scale: a rapid screening
test. Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology
8: 273-278. 1995.
Pratt D and others. Virologic markers of human immunodeficiency
virus type 1 in cerebrospinal fluid of infected children. Journal of
Infectious Diseases 174:288-293. August 1996.
Rottenberg D and others. Abnormal cerebral glucose metabolism
in HIV-1 seropositive subjects with and without dementia. Journal of
Nuclear Medicine 37(7):1133-1141. July 1996.
Sanford J and others. The Sanford Guide to HIV/AIDS Therapy.
Antimicrobial Therapy, Inc., Dallas, 1996.
Toggas S and others. Central nervous system damage produced
by expression of the HIV-1 coat protein gp120 in transgenic mice. Nature
367: 188-193. January 13, 1994.
Tyor W and others. Unifying hypothesis for the pathogenesis
of HIV-associated dementia complex, vacuolar myelopathy, and sensory neuropathy.
Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology
9:379-388. 1995.
Villemagne V and others. Peptide T and glucose metabolism
in AIDS dementia complex. Journal of Nuclear Medicine 37:1177-1180.
July 1996.
Worley J and Price R. Management of neurologic
complications of HIV-1 infection and AIDS. In The Medical Management of
AIDS (third edition). M Sande and P Volberding, eds., W. B. Saunders Company,
Philadelphia, 1992.
Page last updated 17 December 1996
|
