Case Histories - Freeze-drying of Pharmaceuticals and Biopharmaceuticals - page 168

Since solutions were routinely frozen to 501C, it was evident that the
failed product had not been subjected to maximal freeze-concentration
and the residual solution therefore contained an excess of unfrozen
water, causing it to undergo collapse during primary drying. The results
are shown as electron micrographs in Figure 1, and the corresponding
DSC heating traces of the reconstituted solutions in Figure 2. 71
It became clear that the cause of failure resided in an anomalous
formulation of a particular batch of vials. The method of preparing the
solutions relied on the addition of known volumes of buffer, excipients
and protein solution, the latter taken straight from the chromatograph-
ical column, to the vial, but without assaying the protein content. The
protein solution used in the preparation of the failed batch of vials was
later found to have contained an abnormally low amount of protein
that, in turn, gave rise to the depressed T 0 g .
Figure 1 Scanning electron micrographs of freeze-dried Factor VIII preparations, taken
from: (a) a regular production batch with fine, porous structure; and (b) a
failed batch showing extensive cake collapse; bar lines represent 100 mm