The importance of comorbidity in HIV-infected patients over 55: A retrospective case-control study

The Importance of Comorbidity in HIV-infected over 55: A Retrospective Case-Control Study

Patients

Daniel J. Skiest, MD, Dallas, Texas, Eytan Rubinstien, MD, Nina Carley, MD, Hartford, Connecticut, Laura Gioiella, MD, Farmington, Connecticut, Robert Lyons, MD, Hartford, Connecticut

PURPOSE: To study the impact of comorbidity on the course of HIV disease in older patients as compared to a matched cohort of younger patients. METHODS: In a retrospective case-control study, we compared 43 HIV-infected patients >55 years old to a randomly selected cohort of 86 patients t45 years old, matched by date of HIV diagnosis. We collected data on non-HIV-related morbidity (as assessed by the Charlson comorbidity index), initiator of HIV testing, HIV stage at time of HIV diagnosis (TOHD), AIDS defining diagnoses, AIDS-related illnesses (ARI), observed AIDS-free interval, survival, and frequency of HIV-related and unrelated hospitalizations. RESULTS: The older cohort was more likely to have had HIV testing initiated by a health care provider (36 of 36 versus 50 of 66, P = 0.003), and to have acquired HIV from a transfusion (5 of 43 versus 0 of 86, P = O.OOl), had lower CD4 cell counts at TOHD (205 versus 429, P = 0.02), a shorter observed AIDS-free interval (24.0 versus 52.8 months, P = O.ODO2) and a shorter survival (28.2 versus 58.9 months, P = O.ODD2). The older cohort had more HlV-related (13.4 versus 9.2 per 100 patient-months, P = 0.024) and nonHIV-related hospitalizations (12.9 versus 8.1 per 100 patient-months, P = 0.0001). The comorbidity index was significantly higher in the older cohort (0.907 versus 0.198, P = 0.0001) and was a strong predictor of mortality, independent of age group (risk ratio = 1.38 per comorbidity point, P = O.ODO3). CONCLUSIONS: Older HIV-infected patients presented with more advanced disease, which may have been due to lack of HIV awareness in

From the Universrty of Texas Southwestern Medical Center, Dallas, Texas, St. Francis Hosprtal and Medical Center, Mt. Srnai Hospital, Hart ford, Connectrcut, Unrversrty of Connecticut Health Center, Farmington, Connecticut. Requests for reprints should be addressed to Daniel J. Skiest, MD, The Unrversrty of Texas Southwestern Medrcal Center at Dallas, Divrsion of Infectious Diseases, 5323 Harry Hines Blvd., Dallas, Texas 75235-9113. Presented In part at the Second National Conference on Human Retroviruses and Related Infections, Washrngton, DC, 1995, abstract #558. Manuscript submitted March 11, 1996 and accepted in revised form September 23, 1996.

8~1996 by Excerpta All rights reserved.

Medica,

Inc.

this population. Older patients had a shorter observed AIDS-free interval and shorter survival. In addition, they had more HIV- and non--HIVrelated comorbidity. The more rapid course and decreased survival in the elderly may be related to the increase in comorbidity. 0 1996 by Excerpta Medica, Inc. Am J Med. 1996;101:605-

611. s of June 1995, 476,899 individuals with the acsyndrome (AIDS) had Al uired immunodeficiency been reported in the United States.’ Approximately 10% of cases have occurred in persons 50 years of age or older, with persons 55 or older accounting for approximately 6% of cases.’ Transfusion of blood products previously was reported to be the number two risk factor in this population’; however, most of the transfusion-related HIV cases occurredl prior to 1985, when screening of blood products for HIV was initiated in the United States. Older individuals are sexually active and use intravenous drugs, and thus are at signiIicant risk for sexually transmitted and parenterally transmitted diseases, including AIDS.” Previous studies have suggested that the course of HIV is different in patients greater than 55 years of age, with a more rapid progression in this group.4-8 In an attempt to examine the course of HIV/AIDS in older patients and to determine the impact, of nonHIV-related comorbid conditions on morbidity and mortality, we conducted a retrospective cohort study comparing HIV-infected patients >55 years old to a randomly chosen control group <45 years old.

METHODS Study Population The study population was drawn from two hospitals in Hartford, Connecticut: St. Francis Hospital and Medical Center and Mt. Sinai Hospital. St. Prancis Hospital and Medical Center is a 700-b’ed acute care community teaching hospital and Mt. Sinai Hospital is a 300-bed acute care community teaching hospital. Both hospitals are teaching affiliates of the University of Connecticut School of Medicine. The study group (older cohort) consisted of all .HIV seropositive patients older than 55 years of age, who were diagnosed between January 1, 1986 land June 30, 1993. The control group (younger cohort) consisted of a randomly selected cohort of all HIV-pos0002-9343/96/$115.00 PII SOOO2-9343(96)00329-4

605

HIV INFECTION IN OLDER PATIENTS/SKIEST

ET AL

itive patients aged 18 to 45. The control patients were matched to the study patients by date of HIV diagnosis (in 6-month periods), in order to control for any changes in prophylaxis and treatment that may have occurred during the study period and could have impacted morbidity and survival. Two control patients from the same hospital were selected for each study patient.

Study Design Charts were reviewed by one of the authors, using a standardized data collection form with predefined criteria. In addition, if a patient was seen outside of either hospital system, additional information was obtained by contacting outside health care providers, when possible. The following data were collected by review of the inpatient and outpatient medical records and the HIV/AIDS registry, which is maintained by each hospital’s AIDS coordinator: (1) age at time of HIV diagnosis (if this information was unavailable, the date of the patient’s first encounter with the health care system was used.); (2) sex; (3) race; (4) HIV risk factor; (5) Centers for Disease Control and Prevention (CDC) stage at time of diagnosis (1986 definition)g; (6) presence of AIDS by 1987 revised CDC criteria’“; (7) CD4 cell count on presentation (if not available then CD4 count within 6 months of presentation was used); (8) AIDS defining diagnosis (1987 CDC criteria); (9) all AIDS-related illnesses; (10) significant non-HIV-related comorbid conditions; (11) number of hospitalizations after HIV diagnosis and primary reason for hospitalization: HIV-related (defined as either an AIDS-defining illnessaccording to the pre-1993 CDC criteria or a condition secondary to HIV, ie, anemia, diarrhea, etc, as judged by the attending physician or chart reviewer) or non-HIV-related; (12) reason for HIV testing (health care provider initiated versus patient initiated); (13) observed AIDS-free interval (time from HIV diagnosis to AIDS diagnosis); (14) time from HIV diagnosis to death; (15) time from AIDS diagnosis until death; and (16) cause of death as determined by autopsy, death certificate, or if neither available by chart review. The latter was only accepted as cause of death if the cause of death was unequivocally stated in the chart. In order to evaluate the significance of non-HIVrelated comorbid conditions, we used the Charlson Comorbidity Index (excluding AIDS diagnosis), which was developed and validated to control for the potential confounding effects of coexisting morbid conditions, which alter the risk of short-term mortality in longitudinal studies.” This index uses a weighted score based on the relative risk of death within 1 year due to comorbid conditions present upon admission to the hospital. 606

December

1996

The

American

Journal

of Medicine”

Volume

Statistical

Analysis

The chi-square test was used to compare noncontinuous parameters in the two cohorts. The MannWhitney rank-sum test was used to compare continuous data. Kaplan-Meier survival curves were calculated, and the log rank test was used to compare the two cohorts. Comorbid illness was considered as a covariate and incorporated into the survival curves. The risk ratios for survival were calculated using Cox proportional hazards regression analysis, incorporating the presence or absence of comorbid illness. Mean time from HlV to AIDS, HIV to death, and AIDS to death was estimated using Kaplan-Meier curves. Statistical analys:is was performed using SAS Statistical Software package. Data are presented as mean or median ? standard deviation.

RESULTS Over the study period (1986 to 1993) we identified 43 study patients and 86 control patients. Baseline characteristics of the two cohorts are shown in Table I. Follow-up ranged from 1 month tat 90 months, with an overall mean follow-up of 31.1 months. The rank order of the three most prevalent risk factors was similar in the two groups, with intravenous drug use (IDU) being the most prevalent risk factor, followed by male-male sex and heterosexual transmission. IDU accounted for a significantly higher proportion of cases in the younger cohort (62% versus 40%, P = 0.03). There were 5 transfusion-related cases (12%) in the older group compareld with none in the younger group (P = 0.006). The mean age of the transfusion associated cases was 67.3 years (range 61 to 75), which was older than the mean age of the older cohort. In the older group (in those instances where it could be determined) all HIV testing was initiated by a health care provider (36 of 36) versus (50 of 66) in the control group (P = 0.003). The mean CD4 cell count at time of HIV diagnosis was significantly lower in the older cohort (P = 0.02). A higher percentage of patients in the older group presented with symptomatic disease (CDC stage 3 or 4): 18 of 36 (50%) versus 21 of 71 (30%), P = 0.06, as shown in Table I. There were no significant differences in AIDS-defining diagnoses (ADD) between the two cohorts (Table II). Pneumocystis carinii pneumonia (PCP) was the most common ADD in both groups. Esophageal candid&is was the second most prevalent ADD in the older cohort whereas disseminated 1Mycobacterium avium complex (MAC) was the second most prevalent ADD in the younger cohort. Data for all AIDS related illnesses (ARIs) are shown in Table III. PCP was the most common AR1 in both groups. Esophageal candid&is occurred sig101

HIV INFECTION IN OLDER PATIENTWSKIEST

TABLE

I Patient Characteristics Older Cohort

Number of patients Follow-up in months (mean + SD) Age (mean) at TOHD’ 2 SD (range) Male 1%) Race, n (%I Black Caucasian Hispanic Other Ri;;;ctor, n (%I

Younger Cohort

22.9 Z321.8 62 t- 5.3 (55-77) 37 (86%)

35.2 ?25.3 33.1 2 5.5 (19-44) 60 (70%)

24 (56) 12 (28) 6 (14) 1 (2)

t: I:;; 24 (28) 0

17 (40)

* TOHD = time of HIV diagnosis. + Known for 17 patients In older SD = standard deviation, NS = ++ Known for 36 pabents in older +++ Stages 1 and 2 versus 3 and

II AIDS Defining Diagnosis*

Pneumocystis carinii pneumonia Candidiasis, esophageal MAI, disseminated CMV retinitis Tuberculosis (extrapulmonary) HIV Wasting syndrome Cryptococcosis Toxoplasmosis of brain Kaposi’s sarcoma HIV Encephalopathy Isosporiasis, chronic intestinal Total * No significant

differences

were found

cohort and 39 patients not significant. cohort and 71 patients 4.

7 (16) 5 (12) 4 (9)

P 0.037 0.07 NS

0.02 0.03 NS 0.!:6

if21 50 16 20 429

36 (84) 0 7 (16) 205 2 181

o.E3

(58)

(19) (23) 2 391

0.02 0.06+++

1: (50) 4 (11) 14 (39) in younger in younger

: 1

:

1

i

i

:

2:

2:

cohort.

TABLE Ill All Opportunistic

Infections/AIDS Related Illnesses (1987 CDC criteria) Older Younger Cohort, n Cohort, n Pneumocystis carinii pneumonia 10 15 Candidiasis, esophageal* i MAI, disseminatd ; CMV retinitis : : Tuberculosis (extrapulmonary) 2 HIV wasting syndrome 2 : Cryptococcosis 0 Toxoplasmosis of brian :, Kaposi’s sarcoma 1 : HIV Encephalopathy 1 0 Isosporiasis, chronic intestinal 1 CMV, esophageal 2: 35 Total ARls+

diagnoses.

ni6cantly more often in the older cohort, P = 0.01. The younger cohort had a higher number of ARIs per AIDS patient (1.40 versus 1.32) but when length of follow-up after an AIDS diagnosis was considered, this difference was no longer significant (Table IV). There were significant differences in morbidity among the two cohorts as measured by rates of progression to AIDS, and comorbid illnesses (Table IV). There was a greater proportion of patients with AIDS in the older compared with the younger cohort: 22 of 43 (52%) versus 25 of 86 (29%), P = 0.02; however,

5: (70) 5 (7) 16 (23)

cohort.

(1987 CDC criteria) Older Younger Cohort, n Cohort, n 10 13 5 i : 2

for AlDS defining

53 (62) 24 (28) 7 (8)

10 (23)

Male-male sex Heterosexual Transfusion Unknown Initiator of HIV Testing, n (%) Health care provider Patient Unknown CD4 cell count at TOHD, (cells/mm3)+ 2 SD HIV stage at time of diagnosis++ Stage 1, n (%) Stage 2 Stage 3 Stage 4

TABLE

ET AL

Number of patients with any AIDS related illness in older Number of patients with any AIDS related illness In younger * P = 0.01. + ARls = AIDS related illnesses.

cohort cohort

= 22. = 25.

a similar proportion of patients had AIDS diagnosed at their first encounter: 7 (32%) versus 10 (40%). The mean observed AIDS-free interval was shorter in the older cohort: risk ratio = 2.34, P = 0.0002. Tlhe older group had more hospitalizations per patient from all causes (2.8 versus 1.8, P = 0.008); this difference remained when corrected for length of follow-up, (P

December

1996

The American

Journal

of Medicine@

Volume

101

607

HIV INFECTION IN OLDER PATIENTS/SKIEST

TABLE

ET AL

IV Measurements

AIDS diagnosis at first encounter, n (%I* Observed AIDS free interval in months (time from Hospitalizations All hospitalizations/100 patient mo. Non-HIV related (mean I SD) Non-HIV related/l00 patient mo. HIV related mean 2 SD HIV related/100 patient mo. Comorbidity Total comorbidity points Comorbidity points/patient, (mean) AIDS-related illnesses fARIs) Total, n Mean ARls per AIDS patient Mean ARls per AIDS patient-months Median ARls per AIDS patient-months

HIV diagnosis

of Morbidity

until AIDS k SD)

Older Cohort in = 431 7’(32) ’ 24.0 + 21.2

Younger Cohort In = 861 10 cdo, 52.8 ? 28.5

26.3 1.7 k 1.9 12.9 1.1 k 1.9 13.4

17.2 0.9 + 1.4 8.1 0.8 2 1.4 9.2

0.0001 0.007 0.0001

zo7

::98

0.0001

29 1.32 1.17 0.25

35 1.40 1.14 0.11

0.10

P o.oNoso2

o.!Z4

* For 22 patients diagnosed with AIDS In the older cohort and 25 patients in younger cohort.

TABLE

V Survival and Progression

Number deceased, n (%I Interval from HIV diagnosis until death lr SD (months) Interval from AIDS diagnosis until death t SD (months)* Mortality within two months of HIV diagnosis, n 1%) HIV-related death/all deaths+

to AIDS Older Cohort 28.2 2 20.4 6.5 i- 5.6 5 (11.6)

Younger Cohort 20 (23) 58.9 k 27.8 32.2 k 3.1 2 (2.3)

0.006 0.0002

6/l 4

5/l 1

NS

21 (491

P

0.2

+ For 22 patients in older cohort and 25 patients in younger cohort. + Could be determined In 14 of 21 patients in older cohort and 11 of 20 patients in the younger cohort.

= 0.0001). The older cohort had both more non-HIVrelated hospitalizations (12.9 versus 8.1 per 100 patient-months, P = 0.0001) and HIV-related hospitalizations (13.4 versus 9.2 per 100 patient-months, P =

0.024). The higher non-HIV-related comorbidity seen in the older cohort was reflected in the higher rate of hospitalization for non-HIV-related reasons and the higher rate of overall comorbid conditions. The signiflcant comorbid conditions in our two cohorts (based on the Charlson Comorbidity Index, excluding AIDS diagnosis), are listed in the Appendix. The older cohort had a greater number of patients with comorbid conditions: 17 of 43 patients (39.5%) versus 9 of 86 patients (10.5%). They also had a higher weighted comorbidity score: 0.907 versus 0.198 comorbidity points per patient, P = 0.0001. The weighted comorbidity index was found to be a predictor of mortality independent of age group (risk ratio for mortality 1.38 per one point of comorbidity index, P = 0.003). The risk ratio for mortality in the older versus the younger group was 2.17, P = 0.03, independent of comorbidity status. Excluding patients who acquired HIV from a blood transfusion did not alter the comorbidity scores (data not shown). 606

December

1996

The American

Journal

of Medicinez

Volume

The older cohort had a shorter overall survival, although time from AIDS to death did not signiflcantly differ (Table V). Forty-nine percent of patients in the older cohort died versus 23% in the younger cohort. The Kaplan-Meier survival curves are shown in the Figure. The mean survival was shorter in the older group, (P = 0.0002, risk ratio = 3.02; 95% CI 1.63 to 5.59; P = 0.0005)). The presence of IDU as a risk factor for HIV did not alter the survival of the younger cohort, but was associated with a shorter survival in the older cohort by the log rank test (P = 0.004). The cause of death could be determined in 14 patients in the older group and 11 patients in the younger group. The most common AIDSrelated cause of death was PCP in both groups. There was a trend towards mortality within 2 months of HIV diagnosis in the older group: 11.6% versus

2.3%, P = 0.08.

DISCUSSION Although HlV and AIDS occur in older individuals, this topic has received little attention, as has recently been noted.” We compared an older cohort of patients with HlV to a matched younger cohort to examine various clinical and epidemiological features, 101

HIV INFECTION IN OLDER PATIENTS/SKIEST

ET AL

cohort. Using the Charlson Comorbidity Index we demonstrated that the older cohort also had a higher weighted comorbidity index, indicating this group was more ill and was less likely to survive. This is not surprising since patients over 55 years of age would be expected to have developed more comorbid illnesses (such as diabetes mellitus and malignancy) than younger individuals, independent of HIV status; however, this has not been noted previously and its significance has not been assesset17In addition, we found a higher hospitalization rate in the older cohort, for both HIV- and non-HIV-related Figure. Survival after HIV diagnosis. Kaplan-Meier estimate of the conditions. A potential confounding fact.or for the probability of survival after HIV diagnosis, P = 0.0002 by log rank higher comorbidity was the higher number of cases test. acquired via blood transfusion (5 versus; 0) in the older cohort. These patients might have been exincluding the importance of non-HIV-related co- pected to have a higher rate of comorbid illness since morbidity. We found significant differences between they required transfusion; however, excluding these the older and younger cohort with respect to several 5 patients did not alter the comorbidity score of the important parameters including the initiator of HIV older group. The impact of non-HIV-related comortesting, CD4 count at time of HIV diagnosis, ARIs per bid illnesses on progression to AIDS and survival patient-month, number of hospitalizations (both HIV needs to be assessedfurther as it may help explain related and unrelated), the presence of significant the shorter survival and faster progression to AIDS comorbid illnesses, observed AIDS-free interval, and in older patients demonstrated in our study and by others.4-s”4 survival. The older cohort had a shorter observed AIDS-free Intravenous drug use was the most important HIV risk factor for all patients. This differs from previous interval; however, because exact time of .HIV infecseries of patients over 50 years of age, which had tion could not be determined and since the older cofound male-male sex to be the most prevalent risk hort was diagnosed at a later stage of disease, with factor, followed in order by blood transfusion, het- lower CD4 cell counts, we can not exclude lead time erosexual transmission, and IDU.2,7 The finding of bias as a reason for the seemingly faster progression IDU as the most prevalent HIV risk factor may reflect to AIDS. Several series have documented age at time the unique characteristics of our study population, of seroconversion to be an important factsor in time which was largely composed of an inner city popu- to AIDS onset, with older patients consistently havlation. However, recent data from the CDC indicates ing a shorter interval from time of HIV diagnosis to time of AIDS diagnosis in all risk categories.4-s~1”-‘7 that IDU is now the second most common risk factor after male-male sex. l3 In our older cohort, transfu- The faster progression in older individuals may be sion was only the number four risk factor. Indeed, related to CD4 count; however, a more rapid protransfusion of blood products has become an in- gression has also been shown to occur in hemophilcreasingly uncommon HIV risk factor in the United iacs independent of CD4 count.‘s A recent study States since screening of blood products was initi- demonstrated that among hemophiliacs, older paated in 1985. The most recent CDC data indicate that tients have a higher viral load after HIV seroconvertransfusion is now the third most common risk fac- sion.lg Thus, higher viral load in older patients may tor and is only slightly more common than hetero- partially explain the faster progression; however, age remained an independent predictor of progression to sexual transmission. l3 For patients with a diagnosis of AIDS, there was a AIDS in this study. We could not analyze the impact trend towards more ARIs (as measured by median of CD4 counts or viral load on progression to AIDS ARIs per patient-months) in the older cohort when and death, because less than half of the patients incorrected for length of follow-up. This may reflect a cluded in our study had a CD4 count recorded at higher degree of immunosuppression in the older time of diagnosis, and at the time of this study viral group secondary to aging or may reflect other un- load measurements were not routinely performed. A larger proportion of patients in the older cohort known factors. The increase in ARIs may have accounted for the increase in HIV-related hospitaliza- died within 60 days of HIV diagnosis, although this did not reach statistical significance. Other groups tions in the older group. We found a signillcant increase in the number of have reported a higher proportion of AIDS diagnoses non-HIV-related comorbid conditions in the older made within the same month as death in older paDecember

1996

The American

Journal

of Medicine@

Volume

101

609

HIV INFECTION IN OLDER PATIENTS/SKIEST

ET AL

-

tients.” This increased early mortality in the older found to be positive for HIV.27 Furthermore, physicohort may reflect a more rapid progression to AIDS cians may be reluctant to discuss AIDS prevention in this group, older patients presenting later in the strategies and to query their older patients for risk course of their illness, or delay in HIV diagnosis in factors. If clinicians are slow to consider the diagthis age group. The increase in comorbid conditions nosis of AIDS in older patients, it could lead to a in older patients may also be a factor in the increased delay in treatment.“*“’ Finally, most of the AIDS early mortality in this group. awareness and AIDS prevention publicity has been Our finding of significantly shorter survival in the directed at younger individuals, which may lead to a older cohort confirms previous studies in which sur- sense of complacency regarding AIDS in older indivival in HIV-infected patients decreased with in- viduals. creasing age.7~s~Z”~2~ This may be attributable to many As with any such study, ours was limited by its factors including: the lower CD4 count at time of retrospective nature. Although we had strictly prepresentation, which has been noted by others417;the defined criteria for each parameter recorded, we more rapid decline in CD4 cell counts in older pa- can not completely exclude observer bias. Also, tients17; the “natural” decline of the immune system despite efforts to obtain complete follow-up inforthat occurs with aging (specifically decreased T-lym- mation on all patients, there were some follow-up phocyte activity)5,825,2”;increased drug intolerance’5; data in a few patients that could not be obtained. decreased use of antiretroviral agents;7 and the lack Our results may not be readily generalized to other of recognition of HIV/AIDS in this population by populations, which have different characteristics health care providers. 27Recent data indicating that from our population. It would be interesting to older patients have higher viral load after serocon- compare our findings, which are taken from a meversion suggests another reason for the decreased dium-sized city in the northeastern United States survival.1g In addition, the decreased survival of to other cities with large numbers of folder HIVIDUs in the older cohort suggests this may be a pos- infected individuals. sible cofactor in the poorer outcome, although this The incidence of HIV in older patients is likely effect was not observed in the younger group. Fi- to continue to increase due to risk factors other nally, our data suggest the increased incidence of than blood transfusion. Approximately 40,000 new non-HIV-related comorbid conditions negatively im- HIV infections occur annually in the IJnited pacted survival in the older cohort. States.“’ AIDS awareness needs to be increased in Since data on socioeconomic status were were un- older patients in order to decrease risk-taking beavailable, we can not exclude such differences as a haviors. Physicians caring for older patients with possible confounder for the decreased survival in the HIV or AIDS should be aware of the increased likeolder cohort. However, there is no reason to suggest lihood for hospitalization and comorbid condisuch differences, especially since other demographic tions. Health care providers should inquire about factors including race and gender were similar and HIV risk factors in older patients and Icounsel all all patients were seen at one of two inner city hos- their patients about effective HIV prevention stratpitals. Finally, lead time bias is another possible rea- egies. Future efforts at prevention and treatment son for the apparent decreased survival in the older of HIV should include older Americans, and future studies should address the impact of comorbid illcohort. An important finding in our study is that no pa- ness on the decreased survival in older HIV-intients in the older cohort initiated HIV testing, while fected individuals. a majority of patients in the younger group did. This ACKNOWLEDGEMENTS may reflect the lack of awareness of HIV/AIDS in We thank Gayle Madden, RN, for maintenance of the AIDS registry at St. Francis older individuals. Surveys of patients over the age of Hospital and Medical Center, and Tom Carmody, PhD (Unrversity of Texas South50 have indicated a signihcant knowledge deficit western Medical Center at Dallas), for statisbcal analysis. concerning AIDS and its modes of transmission.28~“g In a 1990 study, signiticantly fewer patients over age REFERENCES 1. Centers for Disease Control and Preventron. HIV/AIDS Surveillance Rep. 50, as compared with younger patients, reported 1995;7(11:1-34. talking to their physician about AIDSZ8 Another 2. Ship JA, Wolff A, Selik RM. Epidemiology of acquired immune deficiency study found age ~50 to be positively correlated with syndrome in persons aged 50 years or older. J AIDS. 1991;4:84-88. lack of awareness of HIV infection in patients diag- 3. Stall R, Catania J. AIDS risk behaviors among late middle-aged and elderly nosed with AIDS.“’ AIDS may be diagnosed later in Americans. Arch Intern Med. 1994;154:57-63. older patients, possibly due to lack of recognition by 4. Goedert JJ, Kessler CM, Aledort LM, et al. A prospective study of human rmmunodeficiencyvrrustype 1 infection and the development of AIDS in subjects physicians.7r”7,30In a recent study by El-Sadr et al,27 with hemophrlia. NEJM. 1989;321:1141-1148. 5% of serum samples from patients 260 year of age, 5. Rosenberg PS, Goedert JJ, Biggar RJ. Effect of age at s.eroconversion on who were not known to be HIV seropositive, were the natural AIDS rncubabon distribubon. AIDS. 1994;8:803-810. 610

December

1996

The American

Journal

of Medicine@

Volume

101

HIV INFECTION IN OLDER PATIENTWSKIEST

ET AL

6. Blaxhult A, Granath F, Lidman K, et al. The influence of age on the latency period to AIDS in people infected by HIV through blood transfusion. AIDS.

19. D’Bnen TR, Blattner WA, Waters D, et al. Serum HIV-l RNA levels and time to development of AIDS in the multicenter hemophilia cohort study. JAMA.

1990;4:125-129.

1996;276:105-110.

7. Ferro

S, Salit

IE. HIV infection

in pabents

over

55 years

of age. J AIDS.

1992;5:348-355. 8. Sutin DG, Rose DN, Melvihill M, et al. Survival of elderly patients with transfusion-related acquired mmunodeficiency syndrome. J Am Geriatr Sot. 9. Classlficabon system for human T-lymphotroplc virus type lll/lymphadenopathy-associated virus infectlons. MMWR. 1986;35(20):334-339. 10. Revision of the CDC surveillance case definition for acquired immunodefisyndrome. MMWR. 1987;36US):lS-15s. ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis. 1987;40:373-383. 12. Feldman MD. Sex, AIDS, and the elderly. Arch Intern Med. 1994;154:19-20. 13. Centers for Disease Control and Prevention. H/V/AIDS Surveillance Rep. ciency

11. Charlson

1993;5(4):1-29. Martin

JN, Colford

JM, Ngo L, Tager IB. Effect

of older

age on survival

In

HIV Disease. Infectious Disease Society 33rd Annual Meeting, San Francisco, 1995. Abstract #198. 15. Mariotto AB, Mariotti S, Pezzotti P, et al. Estimation of the acquired immunodeficiency syndrome incubation period in intravenous drug users: a comparison with male homosexuals. Am J Epidemiol. 1992;135:428-437. 16. Carre N, Deveau C, Belanger F, et al. Effect of age and exposure group on the onset of AIDS In heterosexual and homosexual HIV-infected patients. AIDS. 1994;8:797-802. 17. Operskalski EA, Stram

DO, Lee H, et al. Human

immunodeficiency

virus

type 1 infection: relationship of risk group and age to rate of progression to AIDS. J infect DIS. 1995;172:648-655. 18. Phillips AN, Lee CA, Elford J, et al. More Rapid progression to AIDS In older HIV-infected

people:

the role

of CD4’

T-cell

counts.

J AIDS.

history of human of Tcell subsets,

immunoplatelet

21. Bacchetti P, Osmond D, Chaisson RE, et al. Survival patterns of the first 500 patients with AIDS in San Francisco. J Infect Dis. 1988;157:1044-1047. 22. Stehr-Green JK, Holman RC, Mahoney MA. Survival analysis of hemophillaassociated AIDS cases In the US. Am J Pub Health. 1989;79:832-835. 23. Lemp GF, Payne SF, Neal D, et al. Survival trends for patients with AIDS.

1993;41:214-216,

14.

20. Eyster ME, Gail MH, Ballard JO, et al. Natural deficiency virus infections In hemophtllacs: effects counts, and age. Ann Intern Med. 1987;107:1-6.

1991;4:970-

975

JAMA. 1990;263:402-406. 24. Rothenberg R, Woelfel M, Stoneburner immunodeficiency syndrome: experience NEJM. 1987;317:1297-1302. 25. Kendig NE, Adler WH. The implications syndrome for gerontology 1990;45:M77-M81. 26. Saltzman RL, Peterson

research

R, et al. Survival with 5833 cases of the acquired and

geriatric

PK. lmmunodeficlency

with the acquired In New York City. mmunodeficlency

medicine.

J Gerontol.

of the elderly.

Revlnfect

Dls.

1987;9:1127-1139. 27. El-Sadr W, Gettler J. Unrecognized human immunodefictency virus infectlon in the elderly. Arch Intern Med. 1995;155:184-186. 28. Gerber? B, Magulre BT, Coates TJ. Are patients talking to their physicians about AIDS? Am J Pub Health. 1990;80:467-468. 29. Porter K, Wall PG, Evans BG. Factors associated with lack of awareness of HIV infection before diagnosis of AIDS. BMJ. 1993;307:20-23. 30. Hu DJ, Byers R, Fleming PL, Ward JW. Characteristics of persons with late AIDS diagnosis in the United States. Am J Prev Med. 1995;11:1:L4-119. 31. Gerbert B, Bleeker T, Maguire B. Caspers N. Physicians ancl AIDS: sexual risk assessment of patients and wllllngness to treat HIV-Infected patients. Arch intern Med. 1992;7:657-664. 32. AIDS: an expanding AIDS, 1993.

tragedy.

Washington,

DC: National

Cc’mmlssion

on

APPENDIX 1 Significant Comorbid Conditions (excluding AIDS)* Older Cohort, n

* According to the Charlson Comorbldity index.” + (I), (2), (3), or (6) = assigned weight for the listed associated with a higher relative risk of death.

comorbid

condition.

P -

0 2 0 0 1 2 0 1 1 0 1 1 17 0.198

0.000:1

2 2 1 1 0 4 4 5 1 3 2

Myocardial infarct (1F Congestive heart failure (1) Peripheral vascular disease (1) Cerebrovascular disease (1) Dementia (1) Chronic pulmonary disease (1) Ulcer disease (1) Diabetes mellitus (without end organ damage) (1) Moderate-severe renal disease (2) Tumor (without metastases) (2) Moderate-severe liver disease (3) Metastatic tumor (61 Total comorbidity points Mean comorbidity points/patient

3; 0.907 A higher

number

correlates

-

Younger Cohort, n

with more significant

comorbid

disease, -

December

1996

The American

Journal

of Medicine@

Volume

llD1

611