The court relied heavily on Haslar v. Megerman, 104 F.3d 178 (8th Cir. 1997), a previous case dealing with shackling an inmate during medical treatment (albeit not with a pregnant woman). In that case, the 8th circuit upheld the shackling of a "virtually comatose" inmate who later suffered permanent leg damage as a result of the shackles being kept too tight as his leg swelled up. The court justified this result by arguing that
[the shackling policy] serves the legitimate penological goal of preventing inmates . . . from escaping  less secure confines, and is not excessive given that goal. A single armed guard often cannot prevent a determined, unrestrained, and sometimes aggressive inmate from escaping without resort to force. It is eminently reasonable to prevent escape attempts at the outset by restraining hospitalized inmates to their beds . . . .
This case does seem inline with that precedent, but only because both cases use an abstract justification (the flight risk of an inmate -- admittedly reasonable most of the time) in situations where it is woefully inadequate (neither pregnant women in labor nor people in comas represent serious escape risks). Somehow, justifying one bad decision by reference to its similarity to another bad decision is not a major consolation to me.
And, seriously, if there ever was a case that met the nebulous "shock the conscience" standard for a due process violation, this would seem to be it (although I admittedly don't know how the due process clause applies to prison inmates).