[R] survival analysis with right and left censoring and time-dependent covariates

[Berta]

Hi R-users,

I have a data set with a particular type of interval-censored data: right 
and left censored data, but no other type of interval censoring. I need to 
fit a model with covariates to evaluate the influence of them on the 
survival time.

First: I use Surv of library(survival) to create my survival object. I have 
two possibilities:
                                      Surv(c(1,1,NA,3),c(1,NA,2,3),type="interval2")
                                       # 1  1+   2-   3 , I do not [a,b] in 
any case.
                                       Surv(c(1,1,2,3),c(1,2,2,3),c(1,0,2,1),type="interval")
                                          1   1+  2-   3

Second:Given that coxph is not valid for interval-censored type of data, I 
use survreg with covariates. 
survreg(Surv(c(1,1,NA,3),c(1,NA,2,3),type="interval2")  ~covariate)

Third: Now I need to fit a model with time-dependent covariates, but I only 
find examples of how to include time-dependent covariates when using coxph 
without left censoring (data=Rossi.txt,  Jhon Fox), and therfore the "event" 
in Surv is 0 until the event occurrs:

http://cran.r-project.org/doc/contrib/Fox-Companion/appendix-cox-regression.pdf

My question is: How could I fit a survival model that include right-left 
censoring and time-dependent covariates?


Thanks a lot in advance,

Berta.