At the end of follow-up the nodular echo patterns (both CSNP and CLNP) had increased in a statistically significant way ([[chi square].sub.MH] = 114,7; P = 0.0001).
Using the Cox model (Table 4), at univariate analysis many factors were associated at baseline with the evolution in HCC, while at multivariate analysis only AFP: Hr = 1.1 (CI 95%: 1.05-1.2) (P < 0.02), CLNP: Hr = 3.4 (CI 95% = 1.6-6.6) (P = 0.02), age: Hr = 1.05 (CI 95% 1.02-1.1) (P = 0.03), and PH: Hr 2.1 (CI 95%: 1.1-4.1) P = 0.03 were found to be independent predictors of HCC.
Using the Kaplan-Meier method, the US pattern at the end of follow-up showed a cumulative risk % ([+ or -]SE) for HCC of 75% ([+ or -]10%) for patients with CLNP, 23% ([+ or -]10%) with CSNP, 21% ([+ or -]3%) with C pattern, and 0% with the H and BL patterns.
During follow-up, patients who developed more frequently HCC were those with CLNP pattern at enrolment 11/27 (40%; CI 95% 24.4-59.4), in a statistically different manner versus C 35/248 (14%; CI 95% 10.3-14.1) (P < 0.002) and versus CSNP 4/32 (12.5%; CI 95% 5.1-28.2) (P < 0.0001).
Zeolite (CLNP) added into the feed (100 m in size) was assured from Enli Mining Corporation (Izmir, Turkey).
The FAs compositions (% of total FAs) of rainbow trout fed with CLNP in Trial Groups, was given in Table 2.
A minimum value of PUFAs/SFAs ratio is recommended as 0.45 (HMSO, 1994), which is lower than those (1.181 1.499 %) obtained from all Trial Groups treated with CLNP. DHA/EPA ratios ranged from 0.72 to 6.89 in some fresh water fish species (Ozogul et al., 2007) and it was 1.56 in rainbow trout (Saglik Aslan et al., 2007).