History of viral suppression on combination antiretroviral therapy as a predictor of virological failure after a treatment change.

OBJECTIVES: HIV-infected persons experience different patterns of viral suppression after initiating combination antiretroviral therapy (cART). The relationship between such differences and risk of virological failure after starting a new antiretroviral could help with patient monitoring strategies. METHODS: A total of 1827 patients on cART starting at least one new antiretroviral from 1 January 2000 while maintaining a suppressed viral load were included in the analysis. Poisson regression analysis identified factors predictive of virological failure after baseline in addition to traditional demographic variables. Baseline was defined as the date of starting new antiretrovirals. RESULTS: Four hundred and fifty-one patients (24.7%) experienced virological failure, with an incidence rate (IR) of 7.3 per 100 person-years of follow-up (PYFU) [95% confidence interval (CI) 6.7-8.0]. After adjustment, patients who had rebounded in the year prior to baseline had a 2.4-times higher rate of virological failure after baseline (95% CI 1.77-3.26; P<.0001), while there was no increased incidence in patients whose last viral rebound was >3 years prior to baseline [Incidence rate ratio (IRR) 1.06; 95% CI 0.75-1.50; P=0.73] compared with patients who had never virally rebounded. Patients had an 86% (95% CI 1.36-2.55; P<.0001), 53% (95% CI 1.06-2.04; P=0.02) and 5% (95% CI 0.80-1.38; P=0.72) higher virological failure rate after baseline if they were virally suppressed <50%, 50-70% and 70-90% of the time they were on cART prior to baseline, respectively, compared with those virally suppressed >90% of the time. DISCUSSION: Intensive monitoring after a treatment switch is required in patients who have rebounded recently or have a low percentage of time suppressed while on cART. Consideration should be given to increasing the provision of adherence counselling.

Survival of European patients with Kaposi's sarcoma as AIDS-defining condition during the first decade of AIDS. AIDS in Europe Study Group.

OBJECTIVES: To determine whether the survival of patients with Kaposi's sarcoma as first manifestation of AIDS has changed between 1979 and 1989; and to define whether prognosis factors could be identified. PATIENTS AND METHODS: This is a multicentric retrospective cohort study from 52 centers in 17 European countries involving adults AIDS patients diagnosed between 1979 and 1989. Variables such as age, sex, geographical regions, transmission groups, date of Kaposi's sarcoma diagnosis, zidovudine use, CD4+ cell count and concomitant opportunistic infections or AIDS-related malignancies were evaluated by using uni- and multivariable proportional hazard models. Log-rank tests were used to determine which variables were associated with survival. RESULTS: From the 6,546 AIDS patients recruited in the database of the AIDS in Europe Study Group, 1,394 were diagnosed with Kaposi's sarcoma at the time of AIDS diagnosis, from 1979 and 1989. A total of 1,047 Kaposi's sarcoma patients died during the follow-up period. By Kaplan-Meier analyses, the median and mean survival for these Kaposi's sarcoma patients were 17 and 25 months, respectively, with no change over time. However, age, sex (female), geographic region, low CD4+ cell count (< 150 x 10(6)/l) and some opportunistic infections and non-Hodgkin's lymphoma were associated with a poorer prognosis. Zidovudine use, year of diagnosis and risk factor for HIV-1 infection brought no additional information as predictor of mortality. CONCLUSIONS: This study suggests that the survival of patients with Kaposi's sarcoma as first manifestation of AIDS has remained poor during the last decade in contrast with the overall AIDS survival which had significantly improved from a median of 13-18 months during the same period of observation. There is a need for further prospective information to explain the worse prognosis in women and the geographical variations.

Effect of sex, age and transmission category on the progression to AIDS and survival of zidovudine-treated symptomatic patients.

OBJECTIVE: To evaluate the effect of transmission category and demographic, clinical and immunological characteristics on the progression to AIDS and survival of zidovudine-treated patients. DESIGN: Prospective multicentre cohort study of symptomatic non-AIDS patients. SETTING: Eighty-three clinical centres reporting data to the National Zidovudine Registry. PATIENTS: A total of 1468 patients enrolled between July 1987 and January 1991 were analysed. MAIN OUTCOME MEASURES: Three-year AIDS-free survival probability estimates since therapy start. Cox proportional hazards regression analysis was used to identify independent predictors of progression to AIDS and survival. RESULTS: Faster progression was associated with increasing age (8% risk increase for a 5-year increase), low baseline CD4+ count (39% risk increase for 100 x 10(6)/l cells decrease), and zidovudine > 1000 mg/day (20% risk increase compared with < or = 1000 mg/day). Homosexual men had a 33% risk increase compared with other risk groups. The presence of fever and oral candidiasis at enrolment were also independently associated with a higher risk of progression. Differences in the risk of progression were not significant between men and women. Older age, baseline CD4+ count, homosexual behaviour, fever and oral candidiasis were independently associated with a shorter survival. CONCLUSIONS: Our results confirm that age and baseline CD4+ count are independent predictors of progression, but do not provide evidence for differences in clinical outcome between the sexes. The higher risk of progression to AIDS and shorter survival for homosexual men appears to be correlated with the higher risk of developing Kaposi's sarcoma.

Regional survival differences across Europe in HIV-positive people: the EuroSIDA study.

OBJECTIVES: To analyse the survival differences between macro-regions of Europe (northern, central and southern Europe) between 1994 and early 1999, and their possible association with antiretroviral treatment differences. DESIGN: From September 1994 the EuroSIDA study has prospectively followed non-selected HIV-infected people from 50 clinical sites in 18 European countries (n = 7331). METHODS: Cox proportional hazards models were used to compare death rates between regions and to investigate the relationship between treatment usage and regional mortality rates. Kaplan-Meier curves were used to compare survival from the first CD4 lymphocyte count of < 200 x 10(6)/l or < 50 x 10(6)/l. RESULTS: At the time of analysis, the median follow-up was 21 months and there was a total of 1544 deaths. In people with a CD4+ cell count that fell below 200 or 50 x 10(6)/l those from central Europe had a better prognosis compared with those from the two other regions (P < 0.05). Patients from central Europe were more frequently exposed to reverse transcriptase inhibitors and protease inhibitors compared with patients from other regions (P < 0.001). There was a significant difference in risk of death between regions after adjustment for baseline differences in demography, presence of AIDS and level of immunodeficiency (risk of death in central Europe was 37% lower than that in southern Europe (P < 0.0001) and 33% lower than in northern Europe (P < 0.0001)). After adjustment for use of individual antiretroviral agents, intensity of treatment regimen, CD4 lymphocyte count, weight, haemoglobin and development of AIDS as time-dependent covariates, the differences became much smaller (risk in central Europe 13% lower than that in southern Europe (P = 0.071) and 15% lower than in northern Europe (P = 0.054). CONCLUSION: Antiretroviral therapy has been used more aggressively in Europe in recent years, resulting in improved prognosis. In this study we observed that the HIV mortality rate in central Europe was significantly lower than those in northern and southern Europe in the period 1994 to early 1999. This finding appears to be due to the effect on survival of different treatment policies and drug availability in the three regions of Europe during this time period, with central European countries, on average, having introduced more aggressive treatment strategies earlier.

Long-term follow-up of zidovudine therapy in asymptomatic HIV infection: results of a multicenter cohort study. The Italian Zidovudine Evaluation Group.

In 1990 the results of a placebo-controlled study conducted within the AIDS Clinical Trials Group (ACTG 019) showed that in the short term, zidovudine was effective in slowing progression to advanced disease in HIV-infected asymptomatic patients with low CD4 cell counts. More recently, the preliminary results of the Concorde Trial suggested that while the data at 1 year agreed with those of ACTG 019, no sustained clinical benefit was detectable at 3 years for early versus deferred therapy. Therefore, the length of the clinical usefulness of zidovudine in this population is still to be determined. We evaluated the 2-year outcome of zidovudine therapy in asymptomatic patients through the prospective follow-up of a cohort of 936 subjects with low CD4+ (< 500/mm3) counts who strictly satisfied, at enrollment, the inclusion and exclusion criteria of the ACTG 019 trial. The clinical end point of the analysis was the development of AIDS. The majority (72.2%) of the individuals in the cohort acquired HIV infection through intravenous drug use; 26.6% were women. The median baseline CD4 cell count was 308/mm3. At 55 weeks of mean follow-up, the progression rate to AIDS (3.2 events per hundred person-years) appears to be comparable to that already reported at the same mean follow-up time for the ACTG 019 zidovudine-treated asymptomatic patients. After 124 weeks of mean follow-up, the overall rate of progression to AIDS was 5.2 per hundred person-years.(ABSTRACT TRUNCATED AT 250 WORDS)

Virological failure among patients on HAART from across Europe: results from the EuroSIDA study.

OBJECTIVES: To monitor the response to highly active antiretroviral therapy (HAART) over time and the proportions of patients with poor virological control in order to help provide some insight into drug resistance. DESIGN: Analysis of data from the EuroSIDA study; an observational study initiated in 1994 of almost 8500 patients with HIV from across Europe. METHODS: Patients who initiated HAART, and had both a CD4 lymphocyte count and viral load measured in the 3 months prior to starting HAART, were included in analyses. The proportion of patients with a poor virological response (defined as a viral load of > 10,000 copies/ml, using either a single measure or two consecutive measures) at 16 and 48 weeks was determined. Multivariate logistical regression was used to determine the factors associated with a poor virological response at both time points. RESULTS: Median CD4 cell count at starting HAART was 218 cells/mm3 [interquartile range (IQR), 113-327 cells/mm3] and median viral load was 4.36 log10 copies/ml (IQR, 3.57-5.04 log10 copies/ml). At 16 weeks, 16% had a viral load of > 10,000 copies/ml based on a single viral load measure and 10% if the more stringent definition of two consecutive viral loads above this level was used. At 48 weeks these proportions were 19% and 13%, respectively. Compared with patients from Southern Europe, patients from both Central and Northern Europe had approximately half the chance of a poor virological response at 16 weeks (odds ratios 0.53 and 0.47, P = 0.0015 and P < 0.0001, respectively), while at 48 weeks both regions still had approximately a 25% reduced chance of a poor virological response, but this was no longer statistically significant (odds ratio 0.77 and 0.75, P = 0.17 and P = 0.13, respectively). CONCLUSIONS: There were marked difference in virological response to HAART across regions of Europe, which may be partly explained by regional differences in access to HAART and utilisation. If drug resistance is closely related to virological failure, these results may help to provide an early insight into the potential problem of drug resistance across Europe. Continued follow-up is essential to monitor patients with poor virological control.

AIDS dementia complex in the Italian National AIDS Registry: temporal trends (1987-93) and differential incidence according to mode of transmission of HIV-1 infection.

OBJECTIVES: To investigate the occurrence of AIDS dementia complex (ADC) in Italy and its incidence over time, examining possible correlations between this condition and some demographic and immunological variables. DESIGN: Inception cohort. Data collected from the case notification forms of the Italian National AIDS Registry. SUBJECTS: 16813 consecutive AIDS cases reported to the National AIDS Registry from August 1, 1987 through October 31, 1993 were included. STATISTICAL METHODS: All data refer to the time of AIDS diagnosis as reported on the case notification forms. Main analyses of the monthly proportion of ADC cases were by multiple logistic regression. RESULTS: 1364 subjects (8.1%) were reported with a diagnosis of ADC as the first AIDS defining disease, either as the only manifestation or associated with other AIDS defining conditions. At the time of AIDS diagnosis, the observed ADC/AIDS proportion was significantly higher among intravenous drug users (IVDU), 9.1%, compared to heterosexuals, 6.3%, and homo-bisexual men, 5.2%. Simple logistic regression analysis showed a significant (p < 0.0001) quadratic trend in the monthly ADC/AIDS proportion, peaking in March 1990 and decreasing thereafter. Multiple logistic regression, adjusting for month of diagnosis, showed that IVDUs have consistently the highest risk and homo-bisexual men the lowest, although differences tended to decrease with increasing age. Older age, in fact, was highly associated with an increased risk, especially within the homo-bisexual and heterosexual transmission categories. CD4 + cells counts proved to have no significant effect on the risk of progressing to AIDS with ADC rather than with any other AIDS indicative disease. CONCLUSIONS: ADC is a relatively frequent manifestation at diagnosis of AIDS among Italian patients, and particularly in IVDUs. Differences in its occurrence were found according to time of diagnosis, transmission category and age.

A three-year follow-up on a child with low level trisomy 8 mosaicism which was diagnosed prenatally.

We report on a case of trisomy 8 mosaicism detected prenatally in a single clone of amniotic fluid culture, and confirmed on fetal blood and on peripheral lymphocytes after birth. A follow-up was performed over 3 years, showing a clinically normal female with cognitive, neuropsychological, and linguistic development in a normal range.

[A case of central pontine myelinolysis: clinical and electrophysiologic findings]. / Su un caso di mielinolisi pontina centrale: rilievi clinici ed elettrofisiologici.

We described a case of Central Pontine Myelinolysis, typical with regard to pathogenesis (excessive correction of hyponatremia), neuroanatomical damage revealed by imaging techniques (Magnetic Resonance Imaging) and clinical symptomatology. Serially performed BAEPs recordings during patient's hospitalization, showed interesting data on functional recovery which well correlate with the evolution of clinical symptomatology. These data strongly suggest the usefulness of such electrophysiological technique for a prognostic evaluation.

Differential survival of patients with AIDS according to the 1987 and 1993 CDC case definitions.

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 < 0.20 x 10(9)/L [200/microL]), 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.

[Decrease in notifications of AIDS dementia complex in 1989-1990 in Italy: possible role of the early treatment with zidovudine]. / Diminuzione delle notifiche di AIDS Dementia Complex nel periodo 1989-1990 in Italia: possibile ruolo del trattamento precoce con zidovudina.

We evaluated the incidence of AIDS dementia complex (ADC) in groups of patients who acquired infection through different risk behaviours, and attempted to evaluate the possible role of zidovudine (AZT) treatment in preventing or delaying the onset of ADC. The Italian National AIDS Registry was used to study patients with AIDS for whom ADC was reported as an index disease. Relative risk of presenting ADC between different patient categories has been determined. Logistic regression was used to analyse temporal trends in the proportion of AIDS cases presenting with ADC. Of the 6466 cases reported between August '87 and August '90, ADC was seen in 640 (9.9%). I. V. drug addicts had twice the risk (estimated odds ratio: 1.9; 95% confidence interval: 1.5-2.6, p less than 0.001), compared to homo/bisexuals, of presenting with ADC. There is significant evidence (p less than 0.0001) that after a progressive increase in the period '87-'89, it began a definite decrease in the monthly proportion of ADC cases, starting August '89. AZT was introduced in Italy in July 1987 for patients with AIDS or advanced ARC. The incidence of AIDS dementia complex at the moment of AIDS diagnosis in our population of patients, began to decline 24 months after the introduction of systematic AZT treatment in Italy. This could have been due to inhibition of HIV replication in the Central Nervous System among patients who initiated AZT-treatment before developing full-blown AIDS.

Time to virological failure of 3 classes of antiretrovirals after initiation of highly active antiretroviral therapy: results from the EuroSIDA study group.

OBJECTIVE: The purpose of the present study was to determine the prevalence and incidence of virological triple drug-class failure (TCF) and to summarize the clinical outcome for patients who started receiving highly active antiretroviral therapy (HAART). METHODS: The present study is an observational longitudinal study of 3496 treatment-experienced (TE) and treatment-naive (TN) patients monitored from the time they started receiving HAART (baseline) until TCF occurred (as determined on the basis of viral loads), until AIDS was newly diagnosed, or until death. RESULTS: Four hundred forty-five patients (12.7%) had TCF; 370 (16.6%) of 2230 patients were TE, and 75 (5.9%) of 1266 patients were TN. At 6 years after starting HAART, 21.4% of TE and 11.2% of TN patients had TCF (P<.0001). The prevalence of TCF at or after 2002 was 15.5% in TE patients and 4.8% in TN patients. TN patients had a 32% annual increase in the incidence of TCF (95% confidence interval [CI], 14%-54%; P<.0001); at 5 years after starting HAART, the rate was comparable for TE and TN patients (3.3 and 3.4 cases/100 person-years of follow-up [PYFU], respectively). The incidence of new cases of AIDS or death was 2.7 cases/100 PYFU in patients who did not experience TCF and 5.0 cases/100 PYFU in patients who did experience TCF, an estimated 36% increase with each category of TCF (95% CI, 19%-56%; P<.0001). CONCLUSION: The prevalence of TCF was low after patients started receiving HAART, particularly among TN patients. Despite the influx of patients who had started receiving HAART more recently, the prevalence of TCF increased over calendar time. Patients with TCF had a higher incidence of newly diagnosed AIDS or death. Treatment of patients with TCF deserves further investigation.

HIV-related non-Hodgkin's lymphoma among European AIDS patients. AIDS in Europe Study Group. AIDS in Europe Study Group.

The epidemiology of HIV associated non-Hodgkin's lymphoma (NHL) was investigated in 6550 European patients with AIDS. NHL was diagnosed in 3.5% of all patients at the time of the AIDS diagnosis. Although the probability of being diagnosed with NHL at AIDS diagnosis was significantly higher among intravenous drug users than among homosexual men, and was associated with increasing age, the observed incidences of NHL were more strikingly similar than any differences. The rate of developing NHL after a previous AIDS diagnosis was 2.4 per 100 patient years of follow-up, and remained constant during a 5-year follow-up period. While primary brain lymphomas comprised only 9% of NHL diagnosed at the time of AIDS, they comprised 38% of NHL diagnosed after AIDS (p < 0.001). The prognosis for patients with NHL at AIDS diagnosis was poor with a median survival of 5 months. A diagnosis of primary brain lymphoma was uniformly associated with a poor outcome. It is concluded that the probability of developing NHL in late stage HIV infection is lower than previously anticipated from the results of small studies on patients receiving long-term anti-retroviral therapy.

AIDS across Europe, 1994-98: the EuroSIDA study.

BACKGROUND: The clinical presentation of HIV-1 related diseases could have changed after the introduction of highly active antiretroviral treatment (HAART). We aimed to assess changes over time in the incidence of ADIs overall and within CD4 lymphocyte count strata, the relationship with treatment and degree of immunodeficiency at diagnosis of ADIs. METHODS: We did a prospective observational multicentre study of over 7300 patients in 52 European HIV-1 outpatient clinics. Incidence rates per 100 patient-years of observation were calculated. FINDINGS: In total, we recorded 1667 new ADIs; the incidence of ADIs declined from 30.7 per 100 patient-years of observation during 1994 (95% CI 28.0-33.4) to 2.5 per 100 patient-years of observation during 1998 (95% CI 2.0-3.0, p<0.0001, test for trend). Median CD4 lymphocyte count at diagnosis of a new ADI increased from 28 cells/microL to 125 cells/microL between 1994 and 1998 (p<0.0001), yet a steep decline in the rate of ADIs was seen after stratification by latest CD4 lymphocyte count within each year (< or = 50, 51-200, and > 200 cells/microL). Patients on HAART had a lower rate of ADIs than patients not on this treatment within each CD4 lymphocyte count strata. The proportion of ADIs attributable to cytomegalovirus retinitis and Mycobacterium avium complex declined over time (p=0.0058 and 0.0022, respectively), whereas the proportion of diagnoses attributable to non-Hodgkin lymphoma has increased (p<0.0001). In 1994, less than 4% of ADIs were non-Hodgkin lymphoma, in 1998 the proportion was almost 16%. This condition has become one of the most common ADIs in patients on HAART. INTERPRETATION: Our findings lend support to the idea that treatment regimens can lower the incidence of ADIs. The immediate risk of an ADI for a given CD4 lymphocyte count has declined over time and is lower among patients on HAART. Long-term follow-up of patients on combination treatment is essential to monitor the incidence of new and emerging diagnoses.

Influence of age on CD4 cell recovery in human immunodeficiency virus-infected patients receiving highly active antiretroviral therapy: evidence from the EuroSIDA study.

Influence of age on the CD4 cell response to highly active antiretroviral therapy (HAART) was examined in 1956 patients (median age, 37.2 years) in the EuroSIDA study. Median initial CD4 cell count was 192x106 cells/L, follow-up was 31 months, and time to maximum CD4 cell response was 20 months. Age groups were not different for baseline CD4 cell count, baseline human immunodeficiency virus RNA load, or treatment history. CD4 cell increase, stratified by age quartiles, differed during months 3-36 of HAART (P=.023). Maximum CD4 cell increase from start of HAART differed by age group (P=.0003), as did maximum CD4 cell count (P<10-4). Multivariate analysis confirmed the inverse relationship between age and maximum CD4 cell response (P=.023). Time to a CD4 increase of >200x106 cells/L was shorter for patients in the younger age groups (P=.0026), as confirmed by multivariate analysis (P<10-4). Younger age may favor CD4 cell restoration because of preserved thymic function.

Epidemiology of AIDS dementia complex in Europe. AIDS in Europe Study Group.

The aim of the study was to describe the epidemiology of AIDS dementia complex (ADC) in Europe and to assess the possible role of zidovudine therapy in preventing or delaying its occurrence. We used an inception cohort, with data collected retrospectively from patients' clinical records from 52 clinical centers in 17 countries across Europe. The subjects were 6,548 adult people with AIDS consecutively diagnosed from 1979 to 1989. The main outcome measures were codiagnosis of ADC at the time of AIDS diagnosis and ADC-free time after AIDS diagnosis. ADC was reported in 295 patients (4.5%) at the time of AIDS diagnosis and during follow-up in a further 402 of the 5,160 patients (7.8%) who were diagnosed with AIDS based on diseases other than ADC. Whether at the time of AIDS diagnosis or later, the occurrence of ADC was significantly associated with age, transmission category, and CD4+ cell counts. The risk was greater in older patients (14 and 19% greater, at AIDS diagnosis and after, respectively, for a 5-year difference in age), in i.v. drug users than in homosexual and bisexual men (89 and 60% greater, at AIDS diagnosis and after, respectively), and for people with lower CD4+ cell counts (14 and 30% greater for a reduction of 1 on the natural log scale). Risk was almost double for women than for men. A significant reduction, of approximately 40%, was found in the risk of developing ADC after AIDS diagnosis for patients receiving zidovudine therapy, but this effect was present only during the first 18 months of treatment, irrespective of whether treatment began before or after AIDS diagnosis. In conclusion, an increase in the risk of developing ADC either at the time of AIDS diagnosis or thereafter is associated with increasing age, i.v. drug use, and decreased CD4+ cell count. Women tend to have a higher risk of ADC at the time of AIDS diagnosis. Zidovudine therapy appears to have a definite, but time-limited, effect of protecting patients against ADC development after AIDS diagnosis.

Cytomegalovirus retinitis in patients with AIDS in Europe. AIDS in Europe Study Group.

The incidence of cytomegalovirus (CMV) retinitis and risk factors associated with the condition were studied in patients with the acquired immune deficiency syndrome (AIDS) in a multicenter retrospective cohort study of 6458 patients from 52 centers in 17 countries in Europe. Cytomegalovirus retinitis was diagnosed in 154 patients (2.4%) at the time of AIDS diagnosis, the probability of this diagnosis being significantly higher for those with CD4+ cell counts of < 100/mm3 (3.4%) than with counts of 100-200/mm3 (1.3%) or > 200/mm3 (0.8%). The rate of developing CMV retinitis after AIDS diagnosis was 9.4 per 100 patient years of follow-up. Multivariate analysis showed that risk behavior was significantly associated with the risk of developing CMV retinitis: lower for intravenous drug users [relative risk (RR) 0.47] and those engaged in "other risk behavior" (RR 0.58) than for homosexual men. The risk of developing CMV retinitis after AIDS diagnosis was significantly associated with CD4+ cell count at the time of AIDS diagnosis: for counts < 100/mm3 (RR 2.90) and from 100 to 200/mm3 (RR 2.13), there was a higher risk than for counts > 200/mm3. Patients with Pneumocystis carinii pneumonia, toxoplasmosis, or extraocular CMV infection at time of AIDS diagnosis exhibited an increased risk of developing CMV retinitis. Patients treated with zidovudine exhibited an increased rate of CMV retinitis: RR was 1.75 during and 2.87 after the second year of treatment as compared to those who had not received zidovudine. Median survival after CMV retinitis at time of AIDS diagnosis was eight months.

Changing patterns of mortality across Europe in patients infected with HIV-1. EuroSIDA Study Group.

BACKGROUND: The introduction of combination antiretroviral therapy and protease inhibitors has led to reports of falling mortality rates among people infected with HIV-1. We examined the change in these mortality rates of HIV-1-infected patients across Europe during 1994-98, and assessed the extent to which changes can be explained by the use of new therapeutic regimens. METHODS: We analysed data from EuroSIDA, which is a prospective, observational, European, multicentre cohort of 4270 HIV-1-infected patients. We compared death rates in each 6 month period from September, 1994, to March, 1998. FINDINGS: By March, 1998, 1215 patients had died. The mortality rate from March to September, 1995, was 23.3 deaths per 100 person-years of follow-up (95% CI 20.6-26.0), and fell to 4.1 per 100 person-years of follow-up (2.3-5.9) between September, 1997, and March, 1998. From March to September, 1997, the death rate was 65.4 per 100 person-years of follow-up for those on no treatment, 7.5 per 100 person-years of follow-up for patients on dual therapy, and 3.4 per 100 person-years of follow-up for patients on triple-combination therapy. Compared with patients who were followed up from September, 1994, to March, 1995, patients seen between September, 1997, and March, 1998, had a relative hazard of death of 0.16 (0.08-0.32), which rose to 0.90 (0.50-1.64) after adjustment for treatment. INTERPRETATION: Death rates across Europe among patients infected with HIV-1 have been falling since September, 1995, and at the beginning of 1998 were less than a fifth of their previous level. A large proportion of the reduction in mortality could be explained by new treatments or combinations of treatments.