Yes, anytime a bolted or threaded joint is thermally cycled there is a potential for leakage. The heater is normally on-line spalled so that significant thermal cycles are already experienced in this equipment without leaking, typically. The emergency steam that should purge the heater coil is colder than the process temperature and will cause a thermal cycle that the heater experiences without incident, typically. It is true that the heat capacity of steam is much lower than the flushing oil so the potential of a large thermal cycle exists. You indicated that the flow rate of this flushing oil is low, which helps in limiting the magnitude of the thermal cycle.
The alternative to not purging the heater in your example is for the heater to coke up completely. The return bends will have to be cut off so the heater can be hydroblasted to clear the heater coil. The transfer line from the heater to the coke drums is also at risk of significant coking if there is no purge of the heavy hydrocarbons in that piping. My preference in this case is to purge slowly with the cold flushing oil.