adf test and cross-correlation with missing values

Dear List,

I have multiple time series, all of which (excepting 1) have missing
values.  These run for ~30 years, with monthly sampling.  I need to
determine stationarity, and have tried to use the Augmented Dickey-Fuller
test (adf.test), but this cannot handle missing values.  The same problem
occurs when attempting cross-correlation (ccf).

Could someone please suggest any suitable functions in R to check for
stationarity and to look at cross-correlation when NAs are present in a
time series (and also, which packages these would be in) - or, do I have
to interpolate the missing values first in order to perform these tests on
my time series?

Thankyou,

Lillian.