Divalproex (DVS) is a popular drug widely used in various neurological and psychiatric disorders. Commercially, it is a multisource-drug available in different generic equivalents. Incidents of (class II)-recalls have been repeated over the last years due to failure to consistently meet dissolution specifications. Class II recalls are known to be associated with temporary or medically reversible adverse health consequences. This study aimed to evaluate the dissolution profiles, among other quality attributes, of select FDA-approved extended-release DVS products before and after exposure to conditions usually seen as short-lived and insignificant on product stability, such as pharmacy dispensing and patients’ in-use conditions to assess their possible role in the failures observed. Products were stored for 6 weeks in pharmacy vials at 30 °C/75 % RH to simulate patient in-use conditions, for 12 weeks in unsealed HPDE bottles at 25 °C/65 % RH to simulate the pharmacy storage conditions, and for 3 days in open containers at 40 °C/75 % RH for accelerated stability studies. Physicochemical changes were detected by near infrared imaging, Fourier transformed infrared, X-ray powder diffraction and differential scanning calorimetry. All samples were analyzed for in vitro dissolution. Two products were further selected for in vivo study on Beagle dogs before and after storage. The physicochemical characterization tests revealed changes in tablets’ composition and drug crystallinity over time. An improved discriminatory dissolution test was developed and used in this study. The in vitro release testing revealed that short-lived environmental changes at 30 or 25 °C could fail some unit doses and significantly lower the drug release (average reduction among all products was 12.97 ± 11.3 % and 27.48 ± 10.26 %, respectively). Some extended-release products showed a significant increase in the amount of drug dissolved in the first 6 h (early burst) owing to changes in tablet surface morphology and enhanced drug dissolution. In vivo studies showed a decrease in the AUC0-t by overall average of 21.1 % using the non-transformed data, a decrease that mirrored the dissolution results. The study shows that significant changes can occur during routine drug dispensing and patients’ use that might variably impact the stability and quality of commercial bioequivalent unit doses. It is possible that these changes may also contribute to the adverse effects reported on DVS or upon drug switches that were previously attributed to the intersubject variability. The study findings are encouraging to further investigate the effect of such minor excursions on the drug effectiveness during products’ shelf lives especially for narrow therapeutic index drugs.