Later note. There is no paradox, see quote at end of answer.
No it is not. The explanation involves accelerations it is true but special relativity is perfectly capable of handling accelerations. Read the book Special Relativity by AP French, part of some MIT series on physics. I borrow from the book since it is summed up much better than I can say it.
it has been argued by some writers that an explanation of the twin paradox must involve the use of general relativity. The basis of this view is that the phenomena in an accelerated reference frame (including the behavior of a clock attached to such a frame) are regarded in general relativity as being indistinguishable, over a limited region of space, from the phenomena in a frame immersed in a gravitational field. This has been interpreted as meaning that it is impossible to talk about the behavior of accelerated clocks without using general relativity. Certainly the initial formulation of special relativity, although it leads to explicit statements about the rates of clocks moving at constant velocities, does not contain any obvious generalizations about accelerated clocks...... Nevertheless, for any clock that is not damaged by the acceleration, the effects of a trip can be calculated without bringing in the notions of equivalent gravitational fields. Special relativity is quite adequate to the job of predicting the time lost. It had better be, for (as Bondi has facetiously put it) "it is obvious that no theory denying the observability of accelerations could survive a car trip on a bumpy road." And special relativity has amply proved itself to be a more durable theory than this.
And one more on accelerations within special relativity.
Because Einstein developed a whole new theory (his general theory of relativity, published in 1916) based upon the dynamical equivalence of an accelerated laboratory and a laboratory in a gravitational field, it is sometimes stated or implied that special relativity is not competent to deal with accelerated motions. This is a misconception. We can meaningfully discuss a displacement and all its time derivatives within the context of the Lorentz transformations.
EXTRA. Funny I am using quotes from books - I don't do that normally - but I happen to have a few of those things lying around and I feel that a published work is much better than some web site. So as to special relativity not being competent to handle the "paradox" ot whether or not there even is a paradox I quote from A First Course In General Relativity by Schutz (taken, incidentally, from the preliminary discussion about special relativity).
Elementary introductions to SR often try to illustrate the physical differences between Galilean relativity and and SR by posing certain problems called "paradoxes". The commonest ones include the 'twins paradox', the 'pole in the barn paradox' ......The idea is to pose these problems in language that makes predictions of SR seem inconsistent or paradoxical, and then to resolve them by showing that a careful application of the fundamental principles of SR leads to no inconsistencies at all: the paradoxes are apparent, not real, and result invariably from mixing Galilean concepts with modern ones. Unfortunately, the careless student (or the attentive student of a careless teacher) often comes away with the idea that SR does in fact lead to paradoxes. THIS IS PURE NONSENSE. Students should realize that all 'paradoxes' are really mathematically ill-posed problems, that SR is a perfectly consistent picture of spacetime which has been experimentally verified in countless situations in which gravitational effects can be neglected, and that SR forms the framework in which every modern physicist must construct his theories.
I did the caps stuff back there, not the author. He then gives a very detailed discription of the 'paradox' paying particular attention to the coordinate systems involved. It is too long to quote here since I do not have an electronic copy of the book but I suggest you go and get one and flip forward about 30 pages and do some reading. This applies to some of the other people giving answers. Do some reading. it can be very instructive.
Further note to philip. As Schutz analyzes the "paradox" you do not even need to consider accelerations hence the mention of coordinate systems above. He lays out three coordinate systems: one is on the earth, the other is moving away from the earth at some near light speed and the third is moving towards the earth at the same speed. He then has his space traveler merely jump from one to the other and then shows how the differences in age are realized. I suggest you look up the book since there is a diagram and I do not feel like drawing and posting it,