Table 3.
Univariate and multivariate models for prediction of SVR (for all patients treated with PI-triple therapy; n = 131)
| Univariate analysis | Multivariate analyses** | |||
|---|---|---|---|---|
| Odds ratio (95% CI) | Wald p value | Odds ratio (95% CI) | Wald p value | |
| Viral kinetics | ||||
| eRVR vs. no eRVR | 8.875 (3.66, 21.51) | <0.0001 | 8.875 (3.66, 21.51) | <0.0001 |
| PI-TW 4 < LLOD vs. > LLOD | 7.115 (3.04, 16.65) | <0.0001 | n/a*** | |
| PI-TW 4 viral load > LLOQ vs. < LLOQ | 0.049 (0.01, 0.24) | <0.0001 | n/a*** | |
| Baseline demographic parameters | ||||
| Fibrosis | ||||
| Liver Cirrhosis (Ishak 5 + 6) vs. no Liver Cirrhosis (Ishak 1–4) | 0.528 (0.24, 1.18) | 0.118 | 0.858 | |
| Sex | ||||
| Male vs. female | 0.533 (0.239, 1.187) | 0.123 | 0.194 | |
| Age | ||||
| Patients < 60 years vs. ≥ 60 years | 0.894 (0.358, 2.23) | 0.811 | 0.751 | |
| Baseline viral load | ||||
| High viral load (>800.000 IU/ml) vs. Low viral load (<800.000 IU/ml) | 0.679 (0.29, 1.58) | 0.367 | 0.165 | |
| Genotype | ||||
| 1a vs. 1b | 1.565 (0.678, 3.62) | 0.294 | 0.308 | |
| Baseline platelet count | ||||
| Platelets < 100/nl vs. > 100/nl | 0.1 (0.029, 0.341) | <0.0001 | 0.112 (0.032, 0.394) | 0.001 |
| Previous treatment response | ||||
| Non-response vs. other* | 0.313 (0.137, 0.715) | 0.006 | 0.357 (0.147, 0.867) | 0.023 |
Significant calculations (p < 0.05) are printed bold. For absolute numbers see Table 2. *Other (includes treatment naïve, relapsers, unknown response, pretreatment with other interferon than pegylated interferon). **Multivariate analysis was performed twice: first, all shown parameters were included, showing significant results for “eRVR vs. no eRVR”, and “Platelets < 100/nl vs. > 100/nl”, secondly multivariate analysis was conducted only with “baseline demographic parameters” (excluding parameters of viral kinetics), showing significant results for “Platelets < 100/nl vs. > 100/nl”, and additionally “Non-response vs. other”. ***“PI-TW4 < LLOD vs. > LLOD” and “PI-TW 4 viral load > LLOQ vs. < LLOQ” both had to be excluded from the multivariate analysis due to strong collinearity with “eRVR vs. no eRVR”. The width of the confidence intervals might be due to the limited number of cases in our sample. Abbreviations: eRVR Extended rapid virological response, LLOD Lower limit of detection, LLOQ Lower limit of quantification, PI Protease inhibitor, SVR Sustained virological response, TW Treatment week, n/a Not applicable.