Sankara (who does not use the term jivan-mukti except once in the Bhagavad-gita-bhasya, but "if it walks like a duck ...") under Brhad-aranyaka 1.4.7 justifies the need for nididhyasana the third Vedantic process, when the Self had been known (and, eo ipso, liberation attained) by the application of sravana and manana the first two processes, with the consideration that the karma that had started bearing fruits trumps knowledge of the Self. Thus even the knower of the Self needs restrain (niyama) to guard against the functions of the body, mind, and senses. It must follow, theoretically, that there is a danger, however slight, that the state of liberation could be lost, or perhaps that final liberation could be delayed through relapse in identification.  

It is probably worthwhile to remember that jivan-mukti did not mean the same thing to everyone (and not to forget its aspect of freedom from the requirement to perform one's duties). Thus Bhaskara, who argued vehemently against jivan-mukti as Advaitins understood it under Brahma-Sutra 3.4.26-7, claiming that liberation in life was not possible, nevertheless affirmed explicitly a form of jivan-mukti under BS 4.1.14, that is, freedom from the psychological torments such as passion and aversion (śarīra-pāte tu viduṣo muktir avaśyaṁ-bhāvinīti | dvidhā-muktir jīvad-avasthāyāṁ rāga-dveśa-mohaiḥ tad-aṅgaiś ca madādibhir vimuktiḥ | pātottara-kālam ātyantikīti). Since final liberation happens only after death, provided one maintained perfect meditation on Brahman and the performance of one's ritual till death, it must also follow that this kind of jivan-mukti could theoretically be lost as well, by slacking in meditation and ritual. 

All best

Aleksandar