 |
|
|
| |
Survival of HIV-Infected Patients with Compensated Liver Cirrhosis: accelerated death rate; MELD & transient elastometry predict mortality
|
| |
| |
Reported by Jules Levin
CROI 2010 Feb 16-19 SF
Paula Tuma1, Inmaculada Jarrin2, Julia del Amo2, Eugenia Vispo1, Jose Medrano1, Luz Martin-Carbonero1, Pablo Labarga1,
Pablo Barreiro1 and Vincent Soriano1
1Hospital Carlos III and 2Epidemiology National Center, Instituto de Salud Carlos III, Madrid, Spain
ABSTRACT
Background: Since the advent of HAART, liver-related mortality has become the leading cause of non-AIDS death in HIV+ patients. Complications of end-stage liver disease due to chronic hepatitis B and/or C are the main responsible for this observation. The current incidence and predictors of mortality in the subset of co-infected patients with compensated cirrhosis is not well known.
Methods: All HIV+ individuals on regular follow-up at one referral HIV clinic in Madrid who underwent at least one FibroScan evaluation were identified. Liver cirrhosis was defined for transient elastometry values >14.5 KPa. Mortality was checked in clinical records and at the National Death Registry. Predictors of death were examined using multivariate analysis. The prognostic value of 3 different tools (elastometry, Child-Pugh and MELD) for survival was assessed using Cox proportional models.
Results: From a total population of 1706 HIV+ individuals, 194 (11.4%) were cirrhotic and were prospectively followed since October 2004 until December 2008. Median follow-up was 2.35 years (IQR 1.4 to 5.5 years), which corresponded to 435 person-years of follow-up. Overall, 89% of cirrhotics had chronic hepatitis C, 10.3% chronic hepatitis B, 4.6% hepatitis delta, and 4.1% liver disease of other causes or unknown etiology. The overall mortality rate was 5.8 deaths per 100 patient-years.
"control of HIV replication and CD4 reconstitution with HAART might not completely overcome the deleterious impact of HIV infection on survival in HIV-positive cirrhotics"
Multivariate analyses showed that age ≥50 years (hazard ratio (HR) = 4.76; 95%CI 1.66 to 13.59; P=0.004); CD4 counts <200 cells/mL (HR = 3.01; 95%CI 1.26 to 7.23; P =0.03) and detectable plasma HIV-RNA (HR = 3.97; 95%CI 1.53 to 10.27; P =0.005) were associated with mortality.
A baseline MELD score ≥11 (log-rank test, P =0.03) and transient elastometry values >28.75 KPa (log-rank test, P =0.001) were independent predictors of mortality.
Conclusions: The death rate in HIV+ patients with compensated liver cirrhosis in the HAART era is 5.8% per year (which is almost 4-fold higher than in the general HIV population), higher than in previously reported HIV-negative individuals with cirrhosis (3 to 4%) or in non-cirrhotic HIV+ patients in the HAART era (<2%). Factors associated with increased mortality are older age, low CD4 counts, and detectable plasma HIV-RNA. This observation reinforces the current recommendation to provide HAART to all cirrhotic HIV+ patients regardless CD4 counts. Both MELD scores and especially transient elastometry values accurately predict mortality in this population.
To be published in journal AIDS -
Outcome. Deaths occurred in 25 (12.5%) subjects during follow-up, yielding an all-cause mortality rate of 5.8 deaths per 100 patient-years. Deaths were due to liver-related complications in 13 (52%) while 4 were due to cardiovascular events. Other reasons (accidents, cancers, etc) were recorded in 4 subjects. It was unknown in the remaining 4 cases. Mortality rates according to baseline characteristics are shown in Table 2. Considering tertiles of baseline liver stiffness values (14.5-17.7, 17.8-28.5, and 28.6-75), the mortality rate was 4.1, 1.9 and 12.7 deaths per 100 patient-years, respectively. Person-time was different when classifying subjects based on MELD or Child-Plough scores. Mortality rate in patients with MELD scores <11 and ≥11 was 3.9 and 11.7 deaths per 100 patient-years, respectively. Patients classified as Child-Pugh A and B had mortality rates of 5.3 and 10.4 deaths per 100 patient-years, respectively.
Overall, deaths occurred in 7/110 (6.4%) patients with both undetectable plasma HIV-RNA plus CD4 counts above 200 cells/μL, while there were 6/35 (18%) in patients with lower CD4 counts plus detectable viremia (p=0.08).
Hazards of death. Univariate and multivariate Cox regression analyses for risk factors associated with death in the study population are displayed in Table 3. In a non-adjusted analysis, patients with elastometry values ≥28.75 KPa (hazard ratio [HR]: 2.99, 95% CI: 1.17-7.65, p=0.002) had a higher risk of death. When a multivariate analysis corrected for age, gender, CD4 count, plasma HIV-RNA, mode of transmission, and liver disease etiology, a liver stiffness value >28.75 (HR: 3.46, 95% CI: 1.24-9.69, p=0.02) was predictor of mortality. Figure 1a displays the survival by baseline liver stiffness tertiles; again statistical differences in survival only were seen for elastometry values >28.75 KPa (log-rank test p=0.001).
The risk of death stratified by baseline Child-Pugh scores is also displayed in Table 3. In the unadjusted model, the relative risk of death did not differ between patients with Child-Pugh class A or B (HR: 1.85, 95% CI: 0.73-4.66, p=0.20). After adjustments for age, gender, CD4 count, plasma HIV-RNA, mode of transmission, and liver disease etiology, a Child-Pugh score class B tended to be associated with a higher risk of mortality (HR: 2.07, 95%CI: 0.75-5.71, p=0.16) but without statistical significance. In the Kaplan-Meier curve (Figure 1b), this not significant trend in differences between patients with Child-Pugh class A and B was reproduced.
Finally, the risk of death stratified by baseline MELD score is displayed in Table 3. In the multivariate analysis adjusted for age, gender, CD4 count, plasma HIV-RNA, mode of transmission, and liver disease etiology, patients with a MELD score ≥11 had a higher risk of death (HR: 3.85, 95% CI: 1.53-9.66, p=0.004) than those with lower MELD scores. Likewise, survival was significantly lower in patients with MELD ≥11 than in those with lower MELD scores (Figure 1c).
All analyses were repeated using only the 13 liver-related deaths as main end-point, and the results did not change significantly; however, the limited number of hepatic deaths precluded drawn any reasonable conclusion.
|
| |
|
|
|
|
|
