VAR model: How to get bootstrapped 95% prediction interval?

Question

95% prediction interval for the VAR model can be obtained this way (vars library):

library(vars)
data(Canada)

series_mod = window(Canada, start = c(1990, 1), end = c(2000, 4))
mod = VAR(series_mod, p = 2, type = "const")

predict(mod, n.ahead = 4, ci = 0.95)

But what if I would like to get bootstrapped 95% prediction interval? How can I calculate it? Say, Arima in forecast library has parameters for this:

forecast(model, ..., bootstrap = T, npaths = 10000)

But what about VAR model?