ARIMA for multiple series

I want to use the arima class on multiple independent series, ie fit ARIMA models on a collection of time series with the same parameters across the collection. The likelihood is simply the product of likelihods on each serie. I can't just concatenate those series and call the monovariate arima methods because it would create wrong lagged data at the concatenation indices. Is there any workaround with the arima class, or do I need reimplement the full machinery with usual optimisers ? Thank you.