Routine Laboratory Data Can Help Predict Advanced Liver Fibrosis in HIV-HCV Coinfected Patients

Studies have shown that liver disease may progress more rapidly in patients coinfected with HIV and the hepatitis C virus (HCV). This suggests that these patients may warrant more frequent monitoring of fibrosis, but liver biopsies are uncomfortable, expensive, and carry a small risk of complications.

Thus, researchers have attempted to develop non-invasive indicators of fibrosis progression using biochemical markers and various imaging methods.

As reported at the recent ICAAC, Spanish researchers tested a predictive model based on routine laboratory data. The study included a prospective cohort of 296 HIV-HCV coinfected patients with liver biopsy results, randomly divided into an estimation group of 226 (70%) and a validation group of 70 (30%); baseline characteristics of both groups were similar.

They constructed an "HGM-1" index to predict significant liver fibrosis (stages F2 to F4), which incorporated platelet count, aspartate aminotransferase (AST), and blood glucose level. An "HGM-2" index was likewise developed to predict advanced fibrosis (stages F3-F4), incorporating platelet count, international normalized ratio (INR, a measure of blood clotting time), alkaline phosphatase, and AST.

Results

The areas under the receiver operating characteristic curves (AUROCs) of the HGM-1 index were 0.807 for the estimation group and 0.712 for the validation group.

The HGM-1 index predicted the presence of stage F2-F4 fibrosis with 93% certainty above the high cut-off (0.848), but was less able (55% certainty) to predict the absence of significant fibrosis below the low cut-off (0.316).

The HGM-1 AUROCs were not statistically different than the AUROCs obtained for various other fibrosis indices in the same patient cohort, including the Forn's Index, the AST to platelet ratio index (APRI), and FIB-4.

The HGM-2 index predicted the presence of F3-F4 fibrosis with 92% certainty above the high cut-off (0.598) and the absence of F3-F4 fibrosis with 64% certainty below the low cut-off (0.138).

The AUROCs of the HGM-2 index were 0.844 for the estimation group and 0.815 for the validation group.

In the validation group, the HGM-2 AUROC was significantly better than those obtained for APRI (0.724; P = 0.036) or FIB-4 (0.703; P = 0.044).

The diagnostic accuracies of HGM-1 and HGM-2 indexes in the validation group are shown the table below:

Conclusion

The researchers concluded that, "HGM-2 was able to accurately predict advanced liver fibrosis among HIV-HCV coinfected patients and performed better than APRI and FIB-4 for this purpose. HGM-1, Forn's, and FIB-4 were less accurate [in predicting] absence of significant liver fibrosis."

Hosp. Gregorio Marañón, Madrid, Spain.

10/17/06

Reference
J Berenguer, P Miralles, E Alvarez, and others. Identification of Liver Fibrosis in HIV/HCV Coinfected Patients with a Simple Predictive Model Based on Routine Laboratory Data. 46th ICAAC. San Francisco, CA. September 27-30, 2006. Abstract H-1885.

Conference
Main Page