sis and drug treatment ofcytomegalovirus (CMV) retinitis in acquired immunodeficiency syn- drome (AIDS) patients over a 7-year period. The study subjects ...
RACE AND THE TREATMENT OF CYTOMEGALOVIRUS RETINITIS IN A COHORT OF PATIENTS WITH ACQUIRED IMMUNODEFICIENCY SYNDROME Anthony K. Wutoh, PhD, RPh, Julia Hidalgo, ScD, MPH, MSW, Joseph Bareta, MS, Walter Rhee, MA, Robert Beardsley, PhD, and Scott Steidi, MD, DMA Washington, DC, and Baltimore, Maryland
This historical cohort study assessed the impact of race on critical factors in the diagnosis and drug treatment of cytomegalovirus (CMV) retinitis in acquired immunodeficiency syndrome (AIDS) patients over a 7-year period. The study subjects included 194 adult patients with a history of AIDS who were treated for CMV retinitis between September 1987 and September 1994. Abstracted inpatient hospital medical records and a statewide automated AIDS database were the primary sources of data. Patients were assessed for severity of CMV retinitis at diagnosis, time from initial CMV retinitis diagnosis to first treatment, survival from diagnosis of AIDS, and initiation of drug treatment for CMV retinitis. Results indicated a significant difference in the severity of CMV retinitis at diagnosis by race. Patients diagnosed with early disease were more likely to be white, whereas patients diagnosed with severe disease were more likely to be black. There was no difference in the type of CMV retinitis treatment or patient survival time after diagnosis, nor time to treatment once diagnosed by race. These results suggest that differences in survival may not be the result of discrimination against black patients and may be due more likely to practices associated with accessing medical treatment. (J Natl Med Assoc. 1 998;90:21 4-220.) Key words: cytomegalovirus retinitis * acquired immunodeficiency syndrome * race From the Department of Clinical and Administrative Pharmacy, Howard University College of Pharmacy, Washington, DC, and the Center for AIDS Services Planning and Development, AIDS Administration, Maryland Department of Health and Mental Hygiene, the Department of Pharmacy Practice and Science, University of Maryland School of Pharmacy, and the Department of Ophthalmology, University of Maryland Medical School, Baltimore, Maryland. This study was supported by a contract with Roche Pharma Business (Palo Alto, California) and the Health Care Financing Administration (HCFA) Cooperative Agreement no. 18C-99522. Requests for reprints should be addressed to Dr Anthony K. Wutoh, Howard University College of Pharmacy, 2300 Fourth St, NW, Washington, DC 20059. 214
The demographics of acquired immunodeficiency syndrome (AIDS) cases in the United States have shifted over time from a disease among predominantly white, homosexual males to one that afflicts a growing spectrum of racial, ethnic, religious, geographic, and sexual preference groups. For example, by December 1995, there were 13,082 AIDS cases reported in Maryland, 75% of whom were black patients and 21% of whom were female.' While the demographic characteristics of AIDS patients have shifted and become diverse, efforts to identify and treat AIDS in minorities have not been as successful as those seen in white patients.2'3 Cytomegalovirus (CMV) retinitis is the most common form of CMV disease in patients with AIDS.4 While it is more likely that patients will be diagnosed JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 90, NO. 4
RACE & TREATMENT OF CMV RETINITIS
Maryland is unique in that it maintains an automated AIDS information system. The Human Immunodeficiency Virus Information System (HIVIS) links information from the Maryland AIDS Registry public and private health-care claims, vital statistics records, and hospital, long-term care, and ambulatory care records.'2 Patient records are entered into the database when reported by their physicians to the Maryland Department of Health and Mental Hygiene as meeting the current Centers for Disease Control (CDC) case definition of AIDS. Supplemental surveillance of death certificates, clinical laboratory logs, automated hospital discharge summaries, and health insurance records also are
conducted periodically.13'14 The HIVIS database has been described in detail previously.'2'15 This information system, as well as detailed abstracts of patient medical records, were the primary source of data for this study. A historical cohort was developed from HIVIS files to study the natural history of CMV retinitis in a population of AIDS patients in Maryland. Eligibility criteria included: * ,18 years, * diagnosis of AIDS as defined by CDC criteria, * diagnosis of CMV retinitis in a Maryland inpatient facility by a board-certified ophthalmologist, * CMV retinitis diagnosed between September 1987 and September 1994, and * documentation of CMV retinitis treatment in an inpatient medical record. Patients were excluded if they were diagnosed in a federal or state correctional facility since medical records were not available or if there was incomplete CMV retinitis drug treatment history. Treatment histories were considered incomplete if no date of diagnosis of CMV retinitis or if no date of initiation of therapy for CMV retinitis were noted in the medical record. Treatment histories also were considered incomplete if there was more than a 14day gap in recorded treatment. Patients enrolled in CMV retinitis clinical trials also were excluded due to the potential bias of the more stringent monitoring required for clinical trial participation and consistent evaluation of patients participating in trials compared with nonclinical trial patients. Finally, Hispanic patients were excluded due to the small number identified in this cohort. Study patients were identified by applying an ICD-9-CM coding net specific for CMV retinitis to the HIV/AIDS Registry, Medicaid claims, and the Maryland Health Services Cost Review Commission abstracted discharge summary data files. All adult patients diagnosed with AIDS in a Maryland hospital between September 1987 and September 1994 were initially enumerated (n=617). Patients who were diagnosed out-of-state or living in correctional institutions and federal facilities were excluded from the study group due to inaccessibility of medical records. This criteria resulted in the exclusion of 88 patients. One hundred sixty-seven patients with unretrievable medical records also were excluded from the study group after three attempts to obtain the medical record were unsuccessful. Eighty-eight patients were misdiagnosed
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 90, NO. 4
215
as a result of routine screening rather than symptomatic disease, patients with CMV retinitis typically present with blurred vision, visual field defects, or decreased visual acuity. The final outcome of untreated retinitis is usually blindness. Estimates of the incidence of CMV retinitis in the AIDS population range from 20% to 40%.4-6 When allowed to progress untreated, CMV retinitis can rapidly progress to blindness within 6 to 8 weeks.6'7 Although CMV is the most common cause of lifethreatening opportunistic viral infection in patients with AIDS,8 there is no published literature that evaluates the role of access to drug therapy to treatment outcomes for CMV retinitis. During the study period, only two treatments for CMV retinitis were approved by the Food and Drug Administration: intravenous ganciclovir (Cytovene, Roche Laboratories, Nutley, NewJersey) and intravenous foscarnet (Foscavir, Astra USA Inc, Westboro, Massachusetts). Both of these treatments required an initial induction dosing generally requiring hospitalization and a subsequent lifetime maintenance regimen of intravenous therapy.5 Several studies have suggested that black patients may be diagnosed at a more advanced stage of human immunodeficiency virus (HIV) infection than white patients9'10 and that differences in severi.ty at diagnosis of various opportunistic infections may be explanatory of survival differences noted in the literature9"1 This study assessed the impact of race on critical factors in the diagnosis and drug treatment of CMV retinitis in AIDS patients over a 7-year period. A historical cohort study design was used among AIDS patients in Maryland to investigate this question.
MATERIALS AND METHODS
RACE & TREATMENT OF CMV RETINITIS
Table 1. Severity Scale at First Diagnosis of Cytomegalovirus Retinitis Early to Mild Disease Serious Disease Severe Disease 1 -peripheral 1 -dual eye involvement 1 -blindness in either eye (with lesion[s]in zone 1* 2-retinal detachment in either eye (nonsight-threatening) lesion 2-single eye involvement with of either eye) 3-dual eye involvement with lesion in zone 2 and/or zone 3 2-single eye involvement lesions in zones 1, 2, and 3 (with lesion[s] in zone 1 4-single eye involvement (nonsight threatening) 3-dual eye involvement with (lesions in posterior pole) 1, 2, and 3 peripheral nonsight-threatening 3-single eye involvement with 5-dual eye involvement with >25% lesion in zone 2 and/or zone >25% retinal involvement involvement in either eye with lesions in zones 3 in both eyes 4-lesion(s) in anterior pole
*Definitions of zones adapted from reference 16: 'Zone 1 is defined as the area of the retina within 1500 pm from the edge of the optic nerve or within 3000 pm from the center of the fovea (roughly the posterior pole). Zone 2 extended from the edge of zone 1 anteriorly to a circle identified by the vorted vein ampullae. Zone 3 extended anteriorly from zone 2 to the ora serrata. The extent of retinal involvement is defined as the percentage of retinal area involved."
with CMV retinitis as defined by a lack of diagnosis in the inpatient medical record and were excluded from further analysis. Eight patient charts were missing the date of CMV retinitis diagnosis, and 8 other patients either refused treatment or had treatment delayed due to the nonsight-threatening peripheral nature of the CMV lesion. These patients were excluded from analysis. As noted earlier, Hispanic patients (3 patients) and 61 patients who participated in CMV retinitis clinical trials also were excluded from further analysis for the reasons described above. Application of the exclusion criteria did not result in significant differences in the demographic characteristics of the patients selected for the final study population. A medical record abstraction form was developed in conjunction with an ophthalmologist specializing in retinal diseases, AIDS Administration epidemiologists, and a clinical pharmacist specializing in AIDS therapy. Data accessed through medical records abstraction validated much of the information provided by HIVIS and clarified the dates of induction and doses of ganciclovir and foscamet used. Medical records were abstracted by trained accredited records technicians and were reviewed for validity prior to computer entry. All data entered into the computer database were validated manually. Key dates and other data such as date of AIDS diagnosis and patient demographic characteristics
were validated, by comparing the data recorded in the HIV/AIDS Registry with the Maryland Health Services Cost Review Commission data files and Medicaid data records. If 100% agreement was not obtained, the records were flagged and manually reviewed to confirm key variables. To validate mortality data, death dates from the HIV/AIDS Registry were compared with dates recorded from the medical record abstracts of patients who died during hospitalization. Due to the relatively rapid rate in which CMV retinitis can progress to blindness, a maximum period of 7 days between diagnosis and the induction of treatment represented immediate treatment. Three patients (two whites and one black) whose treatment was deferred due to peripheral lesions were excluded from the study. Chi-squared analysis was used to evaluate whether there was a difference in time to first treatment after diagnosis between black and white patients. The mean and median time to treatment for all patients also was determined. Severity of CMV retinitis at diagnosis was assessed from the written ophthalmologist notes in the inpatient medical record. A severity classification system adapted from the Studies of Ocular Complications of AIDS (SOCA) clinical trials severity assessment was used to categorize patients (Table 1).16 Patients were categorized by the severity of disease noted at their earliest diagnosis of CMV retinitis.
216
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 90, NO. 4
RACE & TREATMENT OF CMV RETINITIS
One hundred forty-three patients had ophthalmologic notes in sufficient detail to categorize them by severity at diagnosis while 51 patients did not have sufficient detail in the medical record to classify them by severity of retinitis. A 20% random sample of patients was selected for comparison of severity categorization. In a masked classification, a retinal ophthalmologist and the study director agreed on 90% of the patient severity assignments. Disparate assignments were determined by consensus. The severity classifications were analyzed using the chi-squared test. The data for the time to death were analyzed using the Kaplan-Meier life table method. All median values reported were obtained from KaplanMeier analyses. Comparisons of survival were done using the log-rank test. This method of survival analysis accounts for patients who were still alive or whose data were censored at the end of the study period. This study was approved by the Institutional Review Board of the Maryland Department of Health and Mental Hygiene.
RESULTS A total of 617 adult AIDS patients initially were identified as having a diagnosis of CMV retinitis between September 1987 and September 1994. After exclusion using the study criteria, 194 patients with confirmed CMV retinitis were included in the analyses. Demographics are shown in Table 2. There was no significant difference in the distribution by race, sex, or mode of transmission when the 617 patients identified through ICD-9-CM codes were compared with the 194 patients who met all of the inclusion criteria (Table 3). Patients were categorized as having severe disease if they were blind in at least one eye, had suffered retinal detachment, or had gross CMV involvement in both eyes. Of the patients with sufficient data in their medical charts to determine the severity of CMV retinitis, 24 were diagnosed with early disease, 65 with serious disease, and 54 with severe disease. There was a statistically significant difference in the severity of CMV retinitis by race; white patients were more likely to be diagnosed at an early stage of disease, whereas black patients were more likely to be diagnosed in the most severe stage of CMV retinitis (Table 4). Fourteen of 24 patients diagnosed with early disease were white, while 38 of 54 patients diagnosed with severe disease were black (x2; P=.024). JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 90, NO. 4
Table 2. Patient Demographics
Variable
No. (/o)
Race
Black White Sex Male Female Mode of transmission Male to male Intravenous drug user Heterosexual Blood transfusion Unknown Mean age* (years)
120 (62) 74 (38)
154 (80) 38 (20) 107 (56.6) 60 (31.7) 1 9 (10) 5 (1.7) 5 36
*Median age: 35 years. In a bivariate chi-squared analysis, there was no significant difference in treatment within 7 days of CMV retinitis diagnosis by race (P=.704) (Table 5). Ganciclovir or foscarnet was initiated within 7 days of diagnosis in 64 of 74 white patients and 106 of 120 black patients. The median time to first induction was
