Full text of DOI:10.1093/gerona/57.3.M173

Journal of Gerontology: MEDICAL SCIENCES
Copyright 2002 by The Gerontological Society of America
2002, Vol. 57A, No. 3, M173–M177
Prevalence and Impact of Medical Comorbidity
in Alzheimer’s Disease
P. Murali Doraiswamy,1 Joel Leon,2 Jeffrey L. Cummings,3 Deborah Marin,4 and Peter J. Neumann5
1Departments of Psychiatry and Medicine, Duke University Medical Center, Durham, North Carolina.
2Polisher Research Institute, Philadelphia Geriatric Center, Pennsylvania.
3Departments of Neurology and Psychiatry, UCLA School of Medicine, Los Angeles, California.
4Department of Psychiatry, Mt. Sinai School of Medicine, New York, New York.
5Center for Risk Analysis, Harvard School of Public Health, Boston, Massachusetts.
Background. We examined the prevalence of comorbid medical illnesses in Alzheimer’s disease (AD) patients at
different severity levels. We also examined the effect of cumulative medical comorbidity on cognition and function.
Methods. Analyses of data from 679 AD patients (Mini-Mental State Exam score range 0–30, mean ꢀ SD ꢁ 11.8 ꢀ
8) from 13 sites (four dementia centers assessing outpatients, four managed care organizations, two assisted living facil-
ities, and three nursing homes) prospectively recruited using a stratification approach including dementia severity and
care setting. Medical comorbidity was quantified using the Cumulative Illness Rating Scale-Geriatric.
Results. Across patients, 61% had three or more comorbid medical illnesses. Adjusting for age, gender, race, and care
setting, medical comorbidity increased with dementia severity (mild to moderate, p ꢂ .01; moderate to severe, p ꢂ
.001). Adjusting for age, educational level, gender, race, and care setting, higher medical comorbidity was associated
with greater impairment in cognition (p ꢂ .001) and in self-care (p ꢂ .001).
Conclusions. Despite the limitation of a cross-sectional design, our initial findings suggest that there is a strong asso-
ciation between medical comorbidity and cognitive status in AD. Optimal management of medical illnesses may offer
potential to improve cognition in AD.
ELATIVELY few studies have examined the preva-
lence of comorbid medical conditions in Alzheimer’s
and family caregiver willing to participate in the study. The
community sample reflected consecutively enrolled patients
who met study criteria. The NH and ALF samples consisted
of all current residents who met criteria. Other details of this
sample have been published (3).
R
disease or the interactions between cognition and physical
illness (e.g., 1–2). Cognitive impairment in Alzheimer’s
disease may potentially lead to inaccurate symptom report-
ing, noncompliance, and delayed or inadequate treatment.
Conversely, physical illnesses and medications may poten-
tially impact cognition or neurodegeneration. In this report,
we use data from a multisetting survey of Alzheimer’s dis-
ease patients to examine the prevalence of medical comor-
bidity and its effects on cognition and function.
After investigator training, demographic, clinical, and
cognitive data were collected using interviews with both pa-
tients and the primary family or institutional caregiver. The
Clinical Dementia Rating Scale (CDR) (4) was used to stage
the severity of the dementia using interviews with the patient
and caregiver. For some analyses, the six categories of the
overall CDR score were collapsed into mild (questionable
and mild), moderate, and severe (severe, profound, and
terminal). The Cumulative Illness Rating Scale–Geriatric
(CIRS-G) rates 14 organ systems on a 5-point scale where 0
indicates no problem and 4 indicates severe problems. We
used the three summary measures (total number of endorsed
categories, sum of scores across the endorsed categories, and
the severity index [the total score divided by number of cate-
gories]) that have been previously validated (5,6). At sites,
the CIRS-G was completed by reviewing patient medical
histories, past records, and physical examinations. Physical
examinations were also conducted for those subjects who
had not had one recently. In calculating the CIRS-G sum-
mary measures, the psychiatric illness category is excluded
because in the CIRS-G, Alzheimer’s disease is classified as
a psychiatric illness. The Mini-Mental State Examination
(MMSE) was performed within two weeks of the disease
METHODS
Subjects and Assessments
This cross-sectional survey recruited 679 Alzheimer’s
disease patients and caregivers from 13 sites [four managed
care organizations, four medical centers, three nursing
homes (NHs), and two assisted living facilities (ALFs)] in
nine states. Patients were stratified by care setting and de-
mentia severity. The study was approved by the Institutional
Review Boards at each institution, and written informed
consent was obtained from family caregivers and, where
possible, the patient. Key inclusion criteria were National
Institute of Neurological and Communicative Disorders and
Stroke–Alzheimer’s Disease and Related Disorders Associ-
ation criteria for probable Alzheimer’s disease; at least 50
years of age; receiving care from one of the four settings;
M173

M174
DORAISWAMY ET AL.
staging with the CDR. Mobility and self-care were measured
using subscales of the Health Utilities Index (7).
age, gender, educational level, and race, all three summary
CIRS-G measures were higher for residents in NHs or ALFs
compared to patients in the community (p ꢂ .001). Adjust-
ing for age, gender, race, and care setting, medical comor-
bidity increased with dementia severity (mild to moderate,
p ꢂ .01; moderate to severe, p ꢂ .001). Adjusted CIRS-G
Index and Total Scores were higher in the severe stage com-
pared to the moderate and mild stages (p ꢂ .001). The num-
ber of Endorsed Categories and adjusted total CIRS scores
also increased with age (p ꢂ .001). CIRS-G scores did not
significantly differ between men and women or between
whites and nonwhites.
At least 45% of all patients had musculoskeletal, geni-
tourinary, and/or ear, nose, and throat (ENT) disorders; at
least 30% had vascular/heart diseases; and at least 25% had
lower gastrointestinal tract and/or endocrine/metabolic dis-
orders (Figure 1). When patients were grouped by dementia
severity level, genitourinary disorders were substantially
greater among patients in the severe stage (69%) compared
to those in the mild (48%; p ꢂ .001) and the moderate
stages (48%; p ꢂ .001). ENT disorders were more prevalent
in moderate than in severe stages (p ꢂ .001).
After adjusting for age, gender, education, and care set-
ting, higher medical comorbidity was associated with lower
MMSE scores (p ꢂ .0001). The total explained variance of
the complete model (R2) was 0.34. Although the influence
of medical comorbidity on reducing cognition was partially
mediated through its association with age, education, gen-
der, and care setting (Figure 2), increased medical comor-
bidity showed an independent effect on cognition. After
controlling for age, gender, education, care setting, and cog-
nition, higher medical comorbidity was also associated with
poor self-care (R2 ꢁ .47, p ꢂ .0001), decreased mobility
(R2 ꢁ .25, p ꢂ .0001), and greater incontinence (R2 ꢁ .48,
p ꢂ .0002). The effect of the interaction between CIRS-G
severity index and age on these measures was significant
only for incontinence (p ꢂ .0001).
Statistical Analyses
The present analyses used t tests, chi-square analysis, and
the general linear model (GLM) procedure in SAS. Results
from the least square means from the GLM program were
tested for differences across the comorbidity measures and
for differences in the distribution of specific illness catego-
ries between severity levels. To examine the effect of co-
morbidity on cognition, we ran three GLM models. The first
tested the effect of age, education, and care setting. The sec-
ond model introduced the CIRS-G Severity Index; the third
introduced the interaction between the CIRS-G Severity In-
dex and age. These same models were also run for inconti-
nence, mobility, and self-care, with the addition of cogni-
tion (MMSE) as another independent variable.
RESULTS
Table 1 summarizes the characteristics of the sample. Of
the 679 patients, 354 were in the community (mean age 87 ꢀ
7 years, mean MMSE 16 ꢀ 8), 161 were in ALFs (83 ꢀ 7
years, 9 ꢀ 8), and 164 resided in NHs (77 ꢀ 8 years, 6 ꢀ
7). Fifty-four percent of community patients had mild de-
mentia, whereas 53% of NH patients had severe dementia.
Compared to NH and ALF residents, community patients
were younger (p ꢂ .001), more likely to be male (p ꢂ
.001), to be married (69% vs 26% vs 15%, p ꢂ .001), to
have spouse caregivers (63% vs 20% vs 7%, p ꢂ .001), to
have lower Alzheimer’s disease severity levels (p ꢂ .001),
and to have higher MMSE scores (p ꢂ .001). Compared to
NH residents, residents in ALFs were younger (p ꢂ .001),
had higher educational levels (p ꢂ .001), were more likely
to be married (p ꢂ .001), to have spouse caregivers (p ꢂ
.001), and present lower Alzheimer’s disease severity levels
(p ꢂ .001) and higher MMSE scores (p ꢂ .001).
Unadjusted, on average, Alzheimer’s disease subjects
had more than three conditions (3.3 endorsed categories), a
CIRS-G Severity Index of 1.9, and a CIRS-G Total Score of
5.5. After adjustment for care setting, dementia severity,
DISCUSSION
To our knowledge, this is the first published study to re-
port the stage-specific prevalence of medical conditions in
mild, moderate, and severe stages of Alzheimer’s disease as
well as in patients at different care settings. Several findings
emerged from this study that complement and extend the
prior literature in this area. For example, McCormick and
colleagues (1) used a population-based dementia registry to
compare comorbidity and symptoms reported by 154 clini-
cally diagnosed, probable Alzheimer’s disease cases with two
control groups (92 nondemented subjects with mild cogni-
tive impairments and 129 cognitively intact controls) fre-
quency-matched for age and gender. They reported that
symptoms of cognitive impairment were more frequent in
Alzheimer’s cases. However, complaints related to other
body systems (e.g., gastrointestinal symptoms, joint pain, vi-
sion problems) were less common in Alzheimer’s cases than
in controls even though comorbidity was similar among all
three groups. They concluded that people with Alzheimer’s
disease may underreport common symptoms not related to
cognitive impairment. Our study found that medical comor-
bidity was greatest in older subjects (as expected) and in se-
Table 1. Alzheimer’s Disease Sample Characteristics by Service
Setting (N ꢁ 679)
Community Assisted Living Nursing Home
Patients
Residents
Residents
Characteristic
(n ꢁ 354)
(n ꢁ 161)
(n ꢁ 164)
Mean Age
Gender: Female, %
Race: White, %
77.1 [8.18]
55.6
85.9
82.6 [7.03]
75.8
96.3
86.6 [7.20]
83.5
89.6
Years of Schooling, M [SD] 12.3 [4.09]
12.1 [3.82]
25.5
19.3
9.1 [7.64]
10.5 [4.48]
14.6
6.7
5.9 [6.83]
Married, %
69.2
63.3
15.7 [7.93]
Spouse Caregiver, %
MMSE Score, M [SD]
Dementia Severity, %
Mild
54.0
30.5
15.5
27.9
34.2
37.9
6.1
40.9
53.0
Moderate
Severe
Notes: See text for details of statistical differences between groups. Severity
was determined using the Clinical Dementia Rating Scale. MMSE ꢁ Mini-Men-
tal State Exam.

MEDICAL COMORBIDITY IN ALZHEIMER’S DISEASE
M175
Figure 1. Percentage of all Alzheimer’s disease (AD) patients with specific comorbid disease categories and by AD severity groups adjusted
for care setting, age, and gender.
vere Alzheimer’s disease subjects in NHs. We did not find
any differences by race, gender, or education. In addition,
our study found that ENT, vascular, musculoskeletal, and
genitourinary disorders were the most prevalent disorders.
ENT and genitourinary disorders increased with dementia
severity independent of age. The findings from our study as
well as those of McCormick and colleagues together suggest
that clinicians should have a heightened awareness of non-
cognitive disorders whose symptoms may be underreported
during health maintenance visits of Alzheimer’s disease pa-
tients.
We also found that medical comorbidity was associated
with impaired mobility and incontinence, most likely due to
the high prevalence of musculoskeletal and urinary disor-
Figure 2. Mean Mini-Mental State Exam (MMSE) scores across CIRS-G Severity Index. The figure is derived using the LSMEANS proce-
dure in the SAS general linear model program. CIRS-G ꢁ Cumulative Illness Rating Scale–Geriatric.

M176
DORAISWAMY ET AL.
ders. Medical comorbidity was also independently associ-
ated with lower cognition and self-care levels. These find-
ings add to prior studies that have evaluated the impact of
comorbidities on health care utilization and mortality in
Alzheimer’s disease patients. Van Dijk and colleagues (2)
investigated the relationship between comorbidity and sur-
vival in an 8-year follow-up study of 606 nursing home de-
mentia patients. In this study, Parkinsonism, atrial fibril-
lation, pulmonary infection, and malignancies were powerful
predictors of mortality and doubled the mortality chances.
They also found that more severely demented patients had
greater comorbidity and that comorbidity and dementia se-
verity independently influenced mortality. In another study
of community-dwelling elderly patients (not selected on the
basis of presence of dementia), Shelton and coworkers (8)
found that the presence of two or more comorbidities was
a significant predictor of future hospitalization or emer-
gency department visit. In addition, several studies have
documented that the total cost of managing patients with
Alzheimer’s disease increased significantly over age- and
comorbidity-matched controls without Alzheimer’s disease
(reviewed in 9). These extra costs have included not only
nursing care, but also medical claims for outpatient care, emer-
gency department visits, and hospitalization. Despite such
evidence, underdiagnosis, coding, and reimbursement barri-
ers often result in subjects with Alzheimer’s disease receiv-
ing inadequate care (9).
The present study design cannot unravel the potential rea-
sons for such associations between higher cumulative co-
morbidity and lower cognition. Speculatively, the effect of
medical comorbidity on cognition in Alzheimer’s disease
may be due to direct effect on the brain, adverse effect of
concomitant medications, and/or impaired test-taking abil-
ity. It is also possible that the relatively high prevalence of
vascular disorders, such as hypertension and coronary artery
disease, may have contributed to the cognitive impairments
in this sample. Accumulating epidemiological and neuro-
pathological evidence supports an association between Alz-
heimer’s disease and several vascular risk factors, such as
hypertension, coronary artery disease, diabetes mellitus,
ischemic white matter lesions, and generalized atherosclero-
sis (10). It is speculated that vascular disease may stimulate
the dementia process or affect its clinical expression. It is
also of interest that one randomized controlled trial has
shown that treatment of systolic hypertension can reduce
the incidence of dementia (11). A nested case-control study
of the United Kingdom General Practice Research Database
(comprising 368 practices) found that individuals 50 years
or older who were prescribed a lipid-lowering statin drug
had a substantially lower risk of developing dementia (12).
In addition, the apolipoprotein E4 allele, a risk factor for
Alzheimer’s disease, is also a vascular risk factor and has
recently been associated with left ventricular dysfunction in
Alzheimer’s disease subjects (13). Taken together, these
data argue for a heightened awareness of vascular risk fac-
tors in subjects at risk for dementia as well as in those al-
ready diagnosed with Alzheimer’s disease. Prospective con-
trolled trials are needed to determine if better identification
and treatment of such risk factors may alter the onset or pro-
gression of Alzheimer’s disease.
Limitations of the current data include the cross-sectional
approach, which cannot confirm causality, and the observa-
tion that this sample may not necessarily be representative
of all Alzheimer’s disease patients. Although physical ex-
amination data were available for all subjects, such exams
were not performed for all patients specifically for this
study. Hence, it is possible that incomplete charting may
have led to some inaccuracies in quantifying comorbidity.
In addition, we did not control for the adequacy of treat-
ment, and it is also possible that medications used to treat
the comorbidities may have affected cognition, rather than
the illness per se. It is well known that demented patients
are sensitive to the anticholinergic side effects of a variety
of medications, including some prescribed for cardiovascu-
lar, respiratory, or genitourinary conditions. The strengths
of this study are the inclusion of subjects across severity
levels/care settings, standardized collection of cognitive,
functional, and medical measures, and the relatively large
sample size.
In summary, our preliminary findings suggest that there
is a strong association between medical comorbidity and
cognition in Alzheimer’s disease. Medical comorbidities
must be given consideration in the design and interpretation
of cognitive and functional studies in dementia. The inclu-
sion of patients with representative medical comorbidity in
clinical trials would allow for greater generalizability of ef-
ficacy, safety, and drug interaction profiles of the drugs be-
ing tested. At present, cholinesterase inhibitors are the only
approved treatment to enhance cognition and global func-
tion in Alzheimer’s disease (14). Further studies are war-
ranted to determine whether early recognition and manage-
ment of comorbid conditions, such as hypertension and
coronary disease, may offer potential to further optimize
cognitive and functional abilities in Alzheimer’s disease.
While these physical health problems are not unique to Alz-
heimer’s disease, they may have more functional impact in
adults with Alzheimer’s disease (15). In addition to incorpo-
rating comorbidity as part of dementia cognitive research
studies, the reduction of comorbid physical illness may also
be a legitimate, independent outcome measure to target in
the effectiveness of the care provided to such subjects (15).
The Cumulative Illness Rating Scale–Geriatric, Charlson
Index, and Greenfield Index of Coexistent Disease are some
of the scales available to quantify comorbidity for such pur-
poses (6,7,15). Given that the vast majority of managed care
organizations still lack formal disease management pro-
grams for Alzheimer’s disease, our findings along with
those in the literature suggest a need for directed efforts to
improve medical management for member subjects with
Alzheimer’s disease.
Acknowledgments
This research was supported by a grant from the Outcomes Research
Group, Pfizer Inc., New York, New York. PMD was supported by the Paul
Beeson award from the American Federation for Aging Research. JL and
JLC were also supported by grants from the National Institute on Aging
(U01 AG10317-05S2 and AG10123, respectively). We thank the investiga-
tors at all the sites who contributed patients and effort to this project.
Address correspondence to P. Murali Doraiswamy, MD, Box 3018,
DUMC, Durham, NC 27710. E-mail: dorai001@mc.duke.edu

MEDICAL COMORBIDITY IN ALZHEIMER’S DISEASE
M177
References
screen: identifying elderly persons at risk for hospitalization or emer-
gency department visit. Am J Manag Care. 2000;6(8):925–933.
9. Fillit HM. The pharmacoeconomics of Alzheimer’s disease. Am J
Manag Care. 2000;6(22 Suppl):S1139–S1144.
10. Skoog I. Vascular aspects of Alzheimer’s disease. J Neural Transm
Suppl. 2000;59:37–43.
1. McCormick WC, Kukull WA, Van Belle G, Bowen JD, Teri L, Larson
EB. Symptom patterns and comorbidity in the early stages of Alzhei-
mer’s disease. J Am Geriatr Soc. 1994;42:517–521.
2. Van Dijk PTM, Dippel DW, van der Meulen JH, Habbema JD. Co-
morbidity and its effect on mortality in nursing home patients with
dementia. J Nerv Ment Dis. 1996;184:180–187.
3. Leon J, Neumann PJ, Hermann RC, et al. Health related quality of life
and service utilization in Alzheimer’s disease. Am J Alzheimer Dis.
2000;15(2):94–108.
4. Morris JC. The Clinical Dementia Rating (CDR): current version and
scoring rules. Neurology. 1993;43:2412–2414.
5. Miller MD, Paradis CF, Houck PR, et al. Rating chronic medical illness
burden in gero-psychiatric practice and research: application of the Cu-
mulative Illness Rating Scale. Psychiatry Res. 1992;41:237–248.
6. Parmelee PA, Thuras PD, Katz IR, Powell Lawton M. Validation of
the Cumulative Illness Rating Scale in a geriatric residential popula-
tion. J Am Geriatr Soc. 1995;43:130–137.
11. Black SE, Patterson C, Feightner J. Preventing dementia. Can J Neu-
rol Sci. 2001;28(S1):S56–S66.
12. Jick H, Zornberg GL, Jick SS, Sehshadri S, Drachman DA. Statins and
the risk of dementia. Lancet. 2000;356(9242):1627–1631.
13. Van der Cammen TJ, Verschoor CJ, van Loon CP, et al. Risk of left
ventricular dysfunction in patients with probable Alzheimer’s diease
with APOE4 allele. J Am Geriatr Soc. 1998;46(8):962–967.
14. Doraiswamy PM, Steffens DC. Combination therapy for Alzheimer’s
disease: what are we waiting for? J Am Geriatr Soc. 1998;46:1322–1324.
15. Warshaw G. Assessing physical status in Alzheimer disease research.
Alzheimer Dis Assoc Disord. 1997;11(s6):66–72.
7. Feeny D, Furlong W, Boyle M. Multi-attribute Health Status Classifi-
cation Systems: Health Utilities Index. Pharmacoeconomics. 1995;
7(6):490–502.
Received June 20, 2001
Accepted July 6, 2001
8. Shelton P, Sager MA, Schraeder C. The community assessment risk