Skip to main content

Table 4 False positive error analysis of PK DDI extraction

From: Leveraging syntactic and semantic graph kernels to extract pharmacokinetic drug drug interactions from biomedical literature

Error categories Example
Negation Preincubation of human liver microsomes with dihydralazine in the presence of NADPH resulted in decreases in phenacetin O-deethylase activity (an indicator of P450 1A2 activity) and testosterone 6beta-hydroxylase activity (P450 3A4), but not in diclofenac 4′-hydroxylase activity (P450 2C9), an indication of inactivation of P450s 1A2 and 3A4 during the dihydralazine metabolism.
Relation between drug and its metabolites In HLMs, cisapride was N-dealkylated to norcisapride (NORCIS) and hydroxylated to 3-fluoro-4-hydroxycisapride (3-F-4-OHCIS) and to 4-fluoro-2-hydroxycisapride (4-F-2-OHCIS).
Uncertainty Because HMR1766 is an inhibitor and warfarin a substrate of CYP2C9, the authors studied whether warfarin pharmacokinetics and pharmacodynamics are influenced by HMR1766.
Comparison The inductive effect of CBZ was about 46 % higher than that of OXCZ, a difference that may be of clinical relevance.
Cross-clause in long sentences Coadministration with ketoconazole (which inhibits CYP3A4) decreased the mean apparent oral clearance of quinine significantly (P < .001) by 31 %, whereas coadministration with fluvoxamine (which inhibits CYP1A2 and to some extent CYP2C19) had no significant effect (P > .05) on the mean apparent oral clearance of quinine.
  1. The drug names and important cue words in each example are bolded