DBP7 AG/G vs A
Dihua Xu
December 24, 2015
data source: “ArMorr DBP DNA 12 23 2015.xlsx”
Matched Sample
matched on age (+/- 2 years), albumin (+/- 0.2 mg/dl), catheter status, and diabetes status, dbp(+/-5), sbp(+/-5).
dat.group1 <- subset(dat2, group==1)
dat.group0 <- subset(dat2, group==0)
dat.match.id <- numeric(0)
for(i in 1:nrow(dat.group1)){
temp <- subset(dat.group0, catheter == dat.group1$catheter[i] &
diabetes == dat.group1$diabetes[i] &
Age <= dat.group1$Age[i] + 2 &
Age >= dat.group1$Age[i] - 2 &
alb14 <= dat.group1$alb14[i] + .2 &
alb14 >= dat.group1$alb14[i] - .2 &
#Sex == dat.group1$Sex[i] &
sbp14 <= dat.group1$sbp14[i] + 5 &
sbp14 >= dat.group1$sbp14[i] - 5 &
dbp14 <= dat.group1$dbp14[i] + 5 &
dbp14 >= dat.group1$dbp14[i] - 5
)
if(nrow(temp) != 0){
dat.match.id <- append(dat.match.id, c(dat.group1$MRN[i], temp$MRN))
}
}
dat.match <- subset(dat2, MRN %in% unique(dat.match.id))Baseline information: matched sample
## Variable DBP7.A DBP7.AG.G P.Value
## 1 Age 66.5 + 13.1 (275) 65.4 + 13.4 (285) 0.328
## 2 Sex 163 / 275 (59.3 %) 147 / 285 (51.6 %) 0.081
## 3 Race2 243 / 275 (88.4 %) 116 / 285 (40.7 %) <.001
## 4 Ethnic 40 / 275 (14.5 %) 37 / 285 (13.0 %) 0.679
## 5 diabetes 129 / 275 (46.9 %) 132 / 285 (46.3 %) 0.955
## 6 catheter 195 / 275 (70.9 %) 202 / 285 (70.9 %) >.999
## 7 BMI 28.3 + 7.5 (275) 28.3 + 7.3 (285) 0.987
## 8 sbp14 139.2 + 15.9 (275) 140.2 + 15.7 (285) 0.463
## 9 dbp14 70.4 + 9.4 (275) 70.7 + 9.1 (285) 0.674
## 10 ca14 8.6 + 0.8 (273) 8.5 + 0.8 (284) 0.226
## 11 cr14 5.7 + 2.1 (268) 6.5 + 2.6 (280) <.001
## 12 alk14 90.9 + 55.6 (249) 97.5 + 50.4 (268) 0.155
## 13 alb14 3.6 + 0.4 (275) 3.6 + 0.3 (285) 0.711
## 14 pth14 254.3 + 244.1 (217) 322.0 + 236.0 (217) 0.003
## 15 phos14 4.7 + 1.6 (274) 4.6 + 1.4 (284) 0.342
## 16 wbc14 8.5 + 2.8 (265) 8.5 + 2.9 (277) 0.941
## 17 ferr14 274.6 + 341.1 (257) 292.7 + 310.3 (267) 0.527
## 18 Ivvitd 181 / 275 (65.8 %) 218 / 285 (76.5 %) 0.007Kaplan Meier Curve: matched sample
km.m <- survfit(Surv(fu, Death)~group, data=dat.match)
ggkmTable(km.m, ystratalabs=c("A", "AG/G"), timeby=50, main="DBP 7: A vs. AG/G")
Cox model: matched sample
cox.match.fit <- coxph(Surv(fu, Death)~group+Age+Race2+diabetes+pth14+
ca14+phos14+alb14+catheter+Ivvitd, dat.match)
summary(cox.match.fit)## Call:
## coxph(formula = Surv(fu, Death) ~ group + Age + Race2 + diabetes +
## pth14 + ca14 + phos14 + alb14 + catheter + Ivvitd, data = dat.match)
##
## n= 431, number of events= 50
## (129 observations deleted due to missingness)
##
## coef exp(coef) se(coef) z Pr(>|z|)
## group0 -0.1938943 0.8237450 0.3247054 -0.597 0.55041
## Age 0.0402860 1.0411085 0.0135641 2.970 0.00298 **
## Race2NW 0.3697035 1.4473055 0.3557284 1.039 0.29867
## diabetes0 0.2049118 1.2274169 0.2996748 0.684 0.49411
## pth14 -0.0007699 0.9992304 0.0008223 -0.936 0.34908
## ca14 0.1724692 1.1882353 0.2020738 0.853 0.39338
## phos14 -0.0358190 0.9648149 0.1082293 -0.331 0.74068
## alb14 -0.3225492 0.7243003 0.4281990 -0.753 0.45129
## catheter0 -0.3771625 0.6858047 0.3289984 -1.146 0.25163
## Ivvitd0 -0.3390295 0.7124614 0.3517461 -0.964 0.33512
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## group0 0.8237 1.2140 0.4359 1.557
## Age 1.0411 0.9605 1.0138 1.069
## Race2NW 1.4473 0.6909 0.7207 2.906
## diabetes0 1.2274 0.8147 0.6822 2.208
## pth14 0.9992 1.0008 0.9976 1.001
## ca14 1.1882 0.8416 0.7996 1.766
## phos14 0.9648 1.0365 0.7804 1.193
## alb14 0.7243 1.3806 0.3129 1.676
## catheter0 0.6858 1.4581 0.3599 1.307
## Ivvitd0 0.7125 1.4036 0.3576 1.420
##
## Concordance= 0.677 (se = 0.041 )
## Rsquare= 0.04 (max possible= 0.752 )
## Likelihood ratio test= 17.51 on 10 df, p=0.06379
## Wald test = 16.47 on 10 df, p=0.08703
## Score (logrank) test = 16.62 on 10 df, p=0.08327Logistic model: matched sample
logit.match.fit <- glm(Death~group+Age+Race2+diabetes+pth14+ca14+phos14+alb14+catheter+Ivvitd,
dat.match, family="binomial")
summary(logit.match.fit)##
## Call:
## glm(formula = Death ~ group + Age + Race2 + diabetes + pth14 +
## ca14 + phos14 + alb14 + catheter + Ivvitd, family = "binomial",
## data = dat.match)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9926 -0.5442 -0.4333 -0.3152 2.4809
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -5.0097336 2.3371762 -2.143 0.03207 *
## group0 -0.2304286 0.3524819 -0.654 0.51328
## Age 0.0450038 0.0148201 3.037 0.00239 **
## Race2NW 0.4303361 0.3901387 1.103 0.27001
## diabetes0 0.2145800 0.3226116 0.665 0.50596
## pth14 -0.0008790 0.0008885 -0.989 0.32252
## ca14 0.1955977 0.2171938 0.901 0.36782
## phos14 -0.0394911 0.1157190 -0.341 0.73290
## alb14 -0.3724413 0.4692916 -0.794 0.42741
## catheter0 -0.4056734 0.3527161 -1.150 0.25009
## Ivvitd0 -0.3770357 0.3810761 -0.989 0.32247
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 309.37 on 430 degrees of freedom
## Residual deviance: 291.41 on 420 degrees of freedom
## (129 observations deleted due to missingness)
## AIC: 313.41
##
## Number of Fisher Scoring iterations: 5