HIV positive patients first presenting with an AIDS defining illness: characteristics and survival

Introduction

Previous studies have shown that the early detection of asymptomatic HIV-1 infection may lead to opportunities for the control of the spread of the infection and to interventions that may improve the prognosis for infected people.1 2 3 4 We identified a group of patients who presented with an AIDS defining illness who did not know that they were HIV positive and therefore had not received medical intervention in relation to HIV infection. We studied these patients to gain further insights into the effect of early medical intervention on the outcome of an AIDS defining illness.

Patients and methods

We retrospectively studied 436 patients who had presented to this hospital with an AIDS defining illness between 1 January 1991 and 31 December 1993. The patients were divided into two groups. In group A (339 patients) the time between the diagnosis of HIV seropositivity and the AIDS defining illness was greater than two months (range 2.1-120.2 months, median 46.1 months). In group B (97 patients) the time between a positive result in an HIV test and the AIDS defining illness was less than two months (range 0.0-2.0 months, median 0.0 months). Patients in group B had not been tested for HIV before developing their illness; seven of them received a clinical diagnosis of AIDS before an HIV test gave positive results, but the diagnosis was not confirmed for up to two months as a result of a delay in the completion of diagnostic tests such as culture of Mycobacterium tuberculosis in alveolar lavage specimens. One illness had precipitated the diagnosis of AIDS in all patients. Twelve patients had secondary AIDS defining illnesses, predominantly Kaposi's sarcoma.

The age at diagnosis of the first AIDS defining illness, sex, ethnic origin, country of origin, risk factors for HIV-1 infection, and the time between being diagnosed HIV positive and presentation with the AIDS defining illness was determined for each patient.5 The survival of patients until 31 December 1994 or death before this date was determined. Eight patients in group A and three patients in group B were lost to follow up after their AIDS defining illness. In addition, the following details were noted for the 305 patients presenting with their first AIDS defining illness during 1992 and 1993: the use of zidovudine or prophylaxis against Pneumocystis carinii pneumonia before the illness and haemoglobin concentration, mean red cell volume, neutrophil count, and CD4 positive T lymphocyte count at the time of the illness.

Results

Patients in group A did not differ significantly from patients in group B with respect to their age at presentation with their AIDS defining illness, sex, ethnic origin, the ratio of British citizens to non-British citizens, or risk factors for HIV infection (table I). Patients in group A did not differ from those in group B with respect to their haemoglobin concentration, neutrophil count, or CD4 positive T lymphocyte count (table II).

Extrapulmonary M tuberculosis infection was diagnosed in 23 of the 339 (7%) patients in group A and in 15 of the 97 (15%) patients in group B (P=0.019, Fisher's exact test) (table III). The wasting syndrome, cytomegalovirus infection, and cryptosporidiosis were diagnosed in only three (3%) patients in group B but in 21 (6%), 31 (9%), and 17 (5%) patients, respectively, in group A (table III).

A total of 305 patients presented with their first AIDS defining illness in 1992 and 1993 (236 patients in group A and 69 patients in group B). In group A, 115 (49%) patients were taking zidovudine and 147 (62%) were receiving prophylaxis against P carinii pneumonia before their illness. No patients in group B were receiving either treatment before their defining illness.

A Kaplan-Meier survival analysis was performed for patients presenting with new AIDS diagnoses between 1991 and 1993 (figure). Patients in group B survived longer with AIDS than patients in group A (P=0.0026, log rank analysis). The observed number of deaths during follow up was 209 in group A and 41 in group B. The expected number of deaths under the null hypothesis and with the size of the groups taken into account was 194 and 56 respectively.

Kaplan-Meier survival curves for patients with AIDS presenting early (group A) and late (group B)

• Download figure
• Open in new tab
• Download powerpoint

Discussion

We identified a group of patients who presented with an AIDS defining illness at the same time as being found to be HIV positive; such patients thus develop AIDS before receiving health advice and medical intervention. Over one fifth of patients presenting with new AIDS diagnoses fell into this category between January 1991 and December 1993 despite there being a high profile HIV centre based in the Jefferiss Wing at St Mary's Hospital. A recent study reported that patients who were unaware of whether they were HIV positive up to nine months before being diagnosed as having AIDS were predominantly heterosexual, non-white, and aged between 15 and 24.6 In our study the two groups of patients did not differ in these respects. However, the criteria used to define late presenters were different in the two studies.

The pattern of AIDS defining illnesses differed between the two groups of patients. The incidence of P carinii pneumonia as the defining illness in patients in group A fell from 32% in 1991 to 16% in 1993, probably as a result of changing local prophylaxis guidelines from dapsone and pyrimethamine to co-trimoxazole.7 8 This is in contrast to the finding that the incidence of P carinii pneumonia in group B remained between 38% and 39% from 1991 to 1993. AIDS defining illnesses associated with severe immune deficiency such as the wasting syndrome, cryptosporidiosis, and cytomegalovirus infection were almost exclusively seen in group A.9 Conversely, there was an excess of AIDS defining illnesses associated with lesser degrees of immune deficiency in group B (particularly extrapulmonary M tuberculosis infection.) This may be because patients in group A presented with AIDS at a later stage in the decline of the function of their immune system than did those in group B. This is borne out by the finding that patients in group A who died within six months of presenting with an AIDS defining illness had lower CD4 positive T lymphocyte counts than patients in group A as a whole (P=0.007, Mann-Whitney U test). Paradoxically, the CD4 count at presentation with the others was higher, although not significantly, in patients in group A than in patients in group B (table III). The Concorde study suggested that a higher CD4 count attributable to the use of zidovudine during asymptomatic HIV infection was not associated with a clinical benefit.10

The survival of patients in group B, who presented late, was found to be significantly better than the survival of patients in group A, who presented early. This finding is consistent with the view that the onset of AIDS is delayed in patients who receive early medical intervention in HIV infection.10 11 12 However, the finding suggests that the subsequent survival of these patients when they develop AIDS may be commensurately decreased. We believe that our data are compatible with the results of the Concorde and other studies, in which early intervention with zidovudine delayed the development of AIDS but did not affect survival overall.10 12 13

In conclusion, this study highlights the scale of the public health problem posed by patients presenting with AIDS coincidental with their first positive result in an HIV test. It also contributes to the debate on the effects of medical intervention on survival after an AIDS defining illness has developed.

We thank Dr Valerie Kitchen for insights into the data and Dr Geraldine A Johnson for her editorial assistance.