• Abstract
• Methods
• Results
• Comment
• REFERENCES

• Table 1
• Table 2
• Figure 1
• Figure 2
• Table 3

OBJECTIVE: To evaluate the impact of the 1993 Centers for Disease Control and Prevention (CDC) revised classification system for human immunodeficiency virus and expanded surveillance case definition for acquired immunodeficiency syndrome (AIDS) on the number of cases and on survival of patients with AIDS.

DESIGN: Retrospective analysis of data from a prospective cohort study of patients treated with zidovudine.

PATIENTS: A total of 3515 patients enrolled in the Italian National Registry of Zidovudine-Treated Patients between July 1987 and December 1991 were analyzed.

MAIN OUTCOME MEASURES: Numbers and survival probability estimates (using the Kaplan-Meier method) for patients satisfying the 1993 CDC case definition compared with patients fulfilling the 1987 CDC classification. Multiple regression analysis was also used to analyze the combined effect of independent variables on survival.

RESULTS: According to the new classification system, the number of AIDS cases in the study population would increase by 188%. While the median survival of patients meeting the 1987 definition was 24 months, at the end of 57 months 53% of patients meeting the 1993 definition were still alive. Among the patients meeting the laboratory criteria for AIDS diagnosis using the new definition (CD4⁺ lymphocyte count <0.20x10⁹/L (200/microliters)), the presence of an AIDS-defining illness was a strong independent predictor of death.

CONCLUSIONS: The application of the new definition results in a considerable increase in the number of cases. Survival for patients classified according to the 1993 definition is much longer than for those classified with the 1987 definition. Clinical status plays a major role in predicting survival outcome among patients whose CD4⁺ lymphocyte counts meet the new case definition.

(JAMA. 1994;271:1197-1199)

The acquired immunodeficiency syndrome (AIDS) is a clinical condition characterized by a wide spectrum of opportunistic diseases associated with the impairment of cellular immunity. The first (1981) Centers for Disease Control and Prevention (CDC) [1] AIDS case definition was based on the presence of unexplained opportunistic infection and/or unusual neoplasm in an individual with no recognized cause for immune dysfunction. Major revisions were made in 1985 [2] and 1987 [3] that added additional infections, neoplasms, and other conditions to the list of AIDS-defining illnesses. In the 1987 revision, laboratory evidence of human immunodeficiency virus (HIV) infection was also introduced. In 1993, the CDC further revised the case definition for AIDS. In this most recent version [4], all patients with a CD4⁺ T-lymphocyte count less than 0.20x10 (⁹)/L (200/microliters) (or percentage of CD4⁺ cells less than 14) were included as AIDS cases, irrespective of their clinical status. Three more opportunistic diseases were also introduced as AIDS index diseases: pulmonary tuberculosis, recurrent pneumonia, and invasive cervical cancer.

As a result of the changes in the case definition, the number of reported AIDS cases in the United States increased significantly [5]. However, its impact on survival estimates has not been extensively studied. We therefore undertook a study, using data from the Italian National Registry of Zidovudine-Treated Patients, designed to compare the number of cases and duration of survival of patients fulfilling the 1987 CDC surveillance case definition with those fulfilling the 1993 revised case definition, as well as to compare survival among the categories included in the new classification.

Methods
We used data on all consecutive patients enrolled in the Italian National Registry of Zidovudine-Treated Patients between July 1987 and December 1991. The registry was established at the Istituto Superiore di Sanita, Rome (the Italian National Institute of Health), in 1987, when zidovudine was licensed in Italy, to collect prospective clinical and laboratory data on all zidovudine-treated patients with HIV. Only specialized hospitals were allowed to prescribe the drug, which was distributed free of charge. As part of this system, subjects were evaluated at clinical centers quarterly, and follow-up forms with clinical events and laboratory data were sent to the coordinating center at the Istituto Superiore di Sanita [6]. Zidovudine therapy was initially restricted to patients with AIDS, to symptomatic patients without AIDS-defining illnesses, or to patients with CD4⁺ lymphocyte counts less than 0.20x10⁹/L. In 1989, however, entry criteria were extended to include asymptomatic patients with CD4⁺ lymphocyte counts less than 0.50x10⁹/L.

All analyses were based on the intention-to-treat approach, regardless of therapy discontinuations. To compare the number of cases using the 1987 definition and the 1993 definition, we retrospectively applied the diagnostic criteria for each of the definitions to the entire cohort, using clinical and laboratory data at enrollment of patients in the registry. Subsequent changes in CD4⁺ lymphocyte counts or clinical conditions were not accounted for. Patients were also retrospectively categorized according to the five categories specified by the 1993 classification (A3: asymptomatic patients with a CD4⁺ lymphocyte count less than 0.20x10 (⁹)/L; B3: symptomatic patients with a CD4⁺ lymphocyte count less than 0.20x10⁹/L, but without an AIDS-defining condition; and C1, C2, and C3: patients with AIDS-defining diseases with a CD4⁺ lymphocyte count of 0.50x10⁹/L or greater, between 0.20 and 0.499x10⁹/L, and less than 0.20x10⁹/L, respectively).

Survival was measured from the start of therapy to date of death. If the patient was still alive, survival was calculated to the cutoff date (September 30, 1992). The cutoff date for data analysis was chosen in order to minimize the AIDS case reporting delay. Survival status was double-checked using data from the Italian National AIDS Registry, also based at the Istituto Superiore di Sanita, which collects all the official AIDS case notifications and HIV-related deaths.

We used the Kaplan-Meier product-limit method [7] to compare survival for patients meeting the 1987 case definition with those meeting the 1993 case definition; P values refer to the log rank test. The same method was used to compare survival for the five categories within the 1993 definition.

We also examined the effect of clinical status on survival of patients with a CD4⁺ lymphocyte count less than 0.20x10⁹/L. Since plots of minus the log-log of the estimate (Kaplan-Meier) survival time indicated nonproportionality in the relative survival between groups, we used a nonproportional survival model. Furthermore, comparison of the Weibull, log-normal, and log-logistic models, with respect to fitted and observed (Kaplan-Meier) survival curves as well as log likelihoods, showed the log-logistic survival distribution [8] to be the most appropriate. The model included as independent variables age, sex, continuous CD4⁺ lymphocyte count, and the different clinical categories (asymptomatic (A3), symptomatic but without clinical AIDS-defining conditions (B3), and clinical AIDS (C3)).

Results
Study Population.--A total of 7336 patients were enrolled between July 1987 and December 1991 Table 1. Of those, 1220 were considered to have AIDS according to the 1987 classification (C1, C2, and C3 categories), while 2677 had symptomatic disease, and 3439 were asymptomatic. When the new definition was applied to this same group of patients, the number defined as having AIDS increased to 3515, representing an increase of 188%. Among those included using the new definition, 1305 were previously defined as having symptomatic disease (B3 category), and 990 were defined as asymptomatic patients with a CD4⁺ lymphocyte count less than 0.20x10⁹/L (A3 category). Table 1

Table 1. Numbers of Patients Belonging to the Different Categories of Human Immunodeficiency Virus- Infected Patients

Table 2. Baseline Characteristics of Patients Belonging to the Categories Defined According to the 1993 Centers for Disease Control and Prevention (CDC) Classification System

Figure 1. Survival analysis of patients classified according to the 1987 and the 1993 acquired immunodeficiency syndrome (AIDS) case definitions

Figure 2. Survival analysis for the five categories of patients included in the 1993 acquired immunodeficiency syndrome case definition

Table 3. Fifty Percent Probability of Survival (Months) for a 30-Year-Old Man With a CD4⁺ Lymphocyte Count Less Than 0.20x10⁹/L

In the first quarter of 1993 in the United States, after the application of the new case definition there was an immediate 204% increase in the number of cases (181% attributable to the inclusion of patients belonging to categories A3 and B3). Some of this increase represented accumulated cases with conditions that did not previously meet the case definition [5]. Therefore, it has been estimated that AIDS reports will increase by approximately 75% above previously expected numbers in 1993 and by 10% to 20% in 1994 [10].

In our population of more than 7000 HIV-positive individuals enrolled in a multicenter zidovudine registry, the number of cases meeting the new case definition was 188% greater than the number meeting the 1987 definition. However, the similarity of this value to that observed in the United States is likely to be coincidental and may not represent the increase that would occur in Italy if the definition were applied to the broader universe of persons known to be HIV-positive. In comparing the increases in AIDS cases between different countries [11,12], extreme caution is required since the new definition is dependent on HIV and CD4⁺ lymphocyte testing practices. While data on symptomatic cases may be relatively comparable from country to country, the data on cases in category A3 and, to a lesser extent, category B3 will vary considerably based on local testing practices. In Italy there are fewer anonymous testing sites per 100 000 population than in the United States, and CD4⁺ lymphocyte testing is perhaps less routinely performed, which would tend to result in the detection of fewer A3 and B3 cases.

In addition to differences in testing practices, caution must be used in interpreting the increase we observed because of some unique features of our study population. Our population consisted of patients enrolled in a zidovudine registry. Our sample is therefore probably biased toward a sicker population, which would tend to result in an underestimate of the ratio of cases meeting the new definition vs those meeting the old definition. Nonetheless, the application of the new definition does result in a substantial increase in the number of cases.

The main finding of our study is that the survival estimate of patients using the 1993 case definition is considerably longer than that of patients classified using the 1987 definition. Using the previous definition, the median survival for patients enrolled in the registry was 24 months; using the new definition, 53% were still alive after 57 months. These findings have important implications for health planners as well as for those involved in providing care and counseling for patients with AIDS.

We were also able to examine the independent effect of clinical status on survival, controlling for age, sex, and CD4⁺ lymphocyte counts. We found that among those with CD4⁺ lymphocyte counts less than 0.20x10⁹/L, survival was highly dependent on clinical status. Those who were asymptomatic or were symptomatic but with no AIDS-defining clinical conditions had considerably better survival outcomes than those who had clinical AIDS, suggesting that while CD4⁺ lymphocyte count is a reasonable predictor of duration of survival among homogeneous clinical groups, the presence of a clinical AIDS-defining condition plays a major role.

Although the new case definition may in many ways be more useful than the old definition in identifying HIV-infected individuals who are at high risk of serious opportunistic illnesses [13], our data suggest that recognition must be given to the greater heterogeneity of cases in the new definition. In view of the current public image of AIDS as an illness that results in death within a relatively brief period after its diagnosis, care must be taken in providing appropriate information on prognosis based on the diagnostic category into which a patient is classified and on educating the public concerning the impact of the new definition.

This study was supported by grants from the Italian Ministry of Health, Istituto Superiore di Sanita, Rome, Italy, AIDS research projects 1991 and 1992. We thank Donato Greco, MD, and Giovanni Rezza, MD, for the data from the Italian National AIDS Registry.

REFERENCES
1. Centers for Disease Control. Update on acquired immune deficiency syndrome (AIDS)--United States. MMWR Morb Mortal Wkly Rep. 1982;31:507-514.

2. Centers for Disease Control. Revision of the case definition of acquired immunodeficiency syndrome for national reporting--United States. MMWR Morb Mortal Wkly Rep. 1985;34:373-375.

3. Centers for Disease Control. Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. MMWR Morb Mortal Wkly Rep. 1987;36:3S-14S.

4. Centers for Disease Control and Prevention. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Morb Mortal Wkly Rep. 1992;41(RR-I7):1-19.

5. Centers for Disease Control and Prevention. Impact of the expanded AIDS surveillance case definition on AIDS case reporting--United States, first quarter, 1993. JAMA. 1993;269;2492. Originally published in: MMWR Morb Mortal Wkly Rep. 1993;42:308-310.

6. Vella S, Giuliano M, Pezzotti P, et al. Survival of zidovudine-treated patients with AIDS compared with that of contemporary untreated patients. JAMA. 1992;267:1232-1236.

7. Kaplan EL, Meier P. Non parametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481.

8. Kalbfleish JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York, NY: John Wiley & Sons Inc; 1980:27-35.

9. Ancelle-Park RA. European AIDS definition. Lancet. 1992;339:671.

10. Centers for Disease Control. Projections of the number of persons diagnosed with AIDS and the number of immunosuppressed HIV-infected persons--United States, 1992-1994. MMWR Morb Mortal Wkly Rep. 1992;41(RR-18):1-29.

11. Chene G, Morlat P, Dabis F. Impact of revision of AIDS case definition. Lancet. 1992;339:1298-1299.

12. Sheppard HW, Winkelstein W, Osmond D, Moss AR. Effect of new AIDS case definition on numbers of cases among homosexual and bisexual men in San Francisco. JAMA. 1991;266:2221.

13. Brettle RP, Gore SM, Bird AG, McNeil AJ. Clinical and epidemiological implications of the Centers for Disease Control/World Health Organization reclassification of AIDS cases. AIDS. 1993;7:531-539.

Acquired Immunodeficiency Syndrome; AIDS Serodiagnosis; CD4 Lymphocytes; Classification; HIV Infections; Prognosis; Survival; Zidovudine