Pharmacogenomic biomarkers do not predict response to drotrecogin alfa in patients with severe sepsis

Annals of Intensive Care

Table 4 Secondary analyses for effects of high APACHE II/SAPS II scores on differential treatment effects of IRP A and IRP B on mortality—conditional logistic regression models including AIM panel PCs

Factor/effect	Estimate	SE	Odds Ratio (OR) estimate	OR 95% CI		P value^c
Factor/effect	Estimate	SE	Odds Ratio (OR) estimate	Lower	Upper	P value^c
Analysis for IRP A
Treatment	−0.15	0.258	0.861	0.519	1.427	0.56
IRP A	−0.08	0.278	0.926	0.537	1.597	0.78
IRP A by treatment interaction^a	−0.01	0.535	0.989	0.347	2.819	0.98
High APACHE II/SAPS II Scores	−0.06	0.352	0.940	0.472	1.872	0.86
High APACHE II/SAPS II scores by treatment interaction^a	0.82	0.315	2.273	1.226	4.215	< 0.01
High APACHE II/SAPS II scores by IRP A Interaction^a	0.15	0.329	1.164	0.611	2.218	0.65
IRP A by treatment by high APACHE II/SAPS II Scores	−0.42	0.651	0.659	0.184	2.360	0.52
AIM Panel PCs^b						< 0.01
Analysis for IRP B
Treatment	−0.13	0.251	0.880	0.538	1.438	0.61
IRP B	0.14	0.280	1.153	0.666	1.999	0.61
IRP B by treatment interaction^a	−0.09	0.541	0.912	0.316	2.635	0.87
High APACHE II/SAPS II Scores	0.10	0.344	1.100	0.561	2.157	0.78
High APACHE II/SAPS II Scores by Treatment^a	0.63	0.302	1.869	1.033	3.379	0.04
High APACHE II/SAPS II Scores by IRP B Interaction^a	−0.35	0.338	0.704	0.363	1.366	0.30
IRP B by treatment by high APACHE II/SAPS II Scores	0.09	0.677	1.092	0.290	4.118	0.90
AIM Panel PCs^b						< 0.01

Analysis for IRP A involves 376 discordant matched sets from a total of 637 matched sets. Analysis for IRP B involves 372 discordant matched sets from a total of 634 matched sets
^aThe interaction odds ratio is a ratio of odds ratios
^bA total of 10 AIM Panel PCs have been included which account for 33.9% of the variance in the AIM Panel data for the Matched-INDICATED Primary Analysis Population based on all cohorts
^cP values are from Wald Chi-square tests; the test of the combined AIM Panel PCs is the sum of the individual Wald Chi-square tests