Monitoring serum albumin levels in HIV-positive women over time and prior to the initiation of HAART can predict disease progression and death, independent of traditional markers of HIV disease progression such as CD4 cell count and HIV viral load, according to a US study published in the May 1st edition of the Journal of Acquired Immune Deficiency Syndromes.
The investigators suggest that falling levels of serum albumin should alert doctors to the need to start anti-HIV therapy in women with CD4 only moderate immune damage, and change treatment in women already taking HAART. Further, as serum albumin tests can be performed easily and cheaply, they may be a useful prognostic test in resource limited countries.
In a five year study nearly 2, 000 HIV-positive American women enrolled in the Women’s Interagency HIV Study (WIHS) from five large cities were monitored to see if a decline in serum albumin was predictive of an increased risk of death from an AIDS-related cause. At baseline and then at six monthly intervals CD4 cell count, HIV viral load, and serum albumin levels were measured to see if a decline in serum albumin over time was predictive of HIV disease progression and if the level of serum albumin immediately prior to starting HAART indicated the likely success of anti-HIV therapy.
An earlier study in the same cohort of patients found that a single measurement of serum albumin at baseline was strongly predictive of survival, and an investigation involving over 100 HIV and hepatitis C co-infected men at a London hospital found that low levels of serum albumin were associated with an increased risk of death.
The women recruited to the study had an average age of 36 years at baseline. Over a third of women had had a previous AIDS defining illness, and average viral load was 23,000 copies/mL, and mean CD cell count was 332 cells/mm3. A baseline serum albumin level below 35 mg/L was present in 5.9% of women, between 35 and 40 mg/L in 22%, between 40 and 42 mg/L in 27% and above 42 mg/L in 44%. Levels of serum albumin were only modestly related to traditional markers of HIV disease progression, including CD4 cell count, viral load and body mass index.
Over the course of the study, the proportion of women with an albumin level below 35 mg/L increased from 5.9% to 10.7%. Serum albumin levels fell by an average of 0.44 g/L/y per year in the women over the five years of the study. Furthermore, in the 397 women who died, the fall was significantly higher at 1.54 g/L/y. In addition, in the women who died, falls in albumin levels could be detected well in advance of their death. In the 21 patients who died by year three of the study, albumin levels had fallen by an average of 1.8 g/L/y in the first three years, but by 3.7 g/L during the final year of life.
When the prognostic value of albumin levels over time was modelled, a serum albumin level of below 35 mg/L at baseline had a risk hazard (RH) of death of 4.5 compared to albumin of above 42 mg/L. This RH increased to 10.0 on the time-dependent model (95% CI: 7.4 - 13.6). A serum albumin level of between 35 and 40 mg had a RH of 2.7 (95% CI: 1.9 - 3.7).
Even when other markers of disease progression and HAART were added to the model, a serum albumin level below 35 mg/L and between 35 and 40 mg/L remained strong predictors of mortality. A serum albumin level of below 35 mg/L in women taking HAART had a RH of 7.0 (95% CI:3.6 - 13.5) compared to a RH of 2.0 for albumin between 35 mg/L and 40 mg/L (95% CI: 1.1 - 3.6). At three years of follow-up, only a little over 50% of women starting HAART with albumin levels below 35 mg/L were still alive, compared to over 80% of women with an albumin level between 35 and 40 mg/L and approximately 98% of women with albumin levels over 40 mg/L
Hepatitis C co-infection did not increase the risk of low serum albumin levels and poor prognosis.
The investigators note that “even though serum albumin is not a specific marker of HIV-1 infection, it is one of the strongest independent predictors of mortality.” In women with severe immune damage, and a CD4 cell count below 200 cells/mm3, the risk of death was increased eight-fold in those with serum albumin below 35 mg/L compared to those with a level above 42 mg/L. The investigators also note that the “predictive ability of serum albumin increased during the study period”.
Monitoring serum albumin could be useful in clinical setting. In patients with moderate immune damage (CD4 cell count between 350 and 500 cells/mm3), serum albumin below 40 mg/L could suggest that HAART should be initiated earlier. Similarly, a fall in albumin over time could indicate that a patient already taking HAART should change therapy. “Lastly, because serum albumin is an inexpensive and widely available measurement, it may be particularly useful in developing countries.”
Further information on this website
Albumin levels associated with poor prognosis in men coinfected with HIV and HCV - news story November 2002.
Feldman JG et al. Serum albumin is a powerful predictor of survival among HIV-1 infected women. JAIDS, 33: 66 - 73, 2003.