CART LDH Confounding
Edward
26/7/2021
Checking LDH as Cofounding Variable
LDH represents intial tumor burden and was correlated with SES groupings. It is important to understand if and how this relationship might affect SES impact on cytokines, kynureine metabolites, clinical outcomes, and PROs. Spearman ranked correlations and Chi-squared tests was done on demogrpahic data. ANOVA done on peak cytokines. Only baseline values were used on kynurenine metabolites and PROs across timepoints to run covariant ANOVA of LDH*income.
LDH as Independent Variable on Demographics, Cytokines,
#Demographics
pcAge<- cor.test(CART15data$Age, CART15data$LogLDH, method=c('pearson'))
pcEd<- cor.test(CART15data$EducationLevelCoded, CART15data$LogLDH, method=c('pearson'))
pclines<- cor.test(CART15data$lines, CART15data$LogLDH, method=c('pearson'))
pcIncome<- t.test(CART15data$LogLDH ~CART15data$IncomeDi)
pcCR<- cor.test(CART15data$LogLDH, CART15data$day28coded, method=c('pearson'))
pcCRSdays<- cor.test(CART15data$`Day to CRS`, CART15data$LogLDH, method=c('pearson'))
pcNTXdays<- cor.test(CART15data$`Day to NTX`, CART15data$LogLDH, method=c('pearson'))
pcCRSgrade<- cor.test(CART15data$`Max Grade CRS`, CART15data$LogLDH, method=c('pearson'))
pcNTXgrade<- cor.test(CART15data$`Max Grade NTX`, CART15data$LogLDH, method=c('pearson'))
pcCRS <-t.test(CART15data$LogLDH ~ CART15data$`CRS (Y=1 /N =0)`)
pcNTX <- t.test(CART15data$LogLDH ~CART15data$`NTX (yes=1, no =0)`)
data.frame(pcAge$p.value, pcEd$p.value, pcIncome$p.value, pclines$p.value, pcCR$p.value, pcCRS$p.value, pcCRSdays$p.value, pcCRSgrade$p.value, pcNTX$p.value, pcNTXgrade$p.value, pcNTXdays$p.value)ggboxplot(CART15data, x = "CRS (Y=1 /N =0)", y = "LogLDH",
color = "CRS (Y=1 /N =0)", palette = c("#00AFBB", "#E7B800"),
ylab = "LogLDH", xlab = "CRS (Y=1 /N =0)")
ggboxplot(CART15data, x = "NTX (yes=1, no =0)", y = "LogLDH",
color = "NTX (yes=1, no =0)", palette = c("#00AFBB", "#E7B800"),
ylab = "LogLDH", xlab = "NTX (yes=1, no =0)")
ggboxplot(CART15data, x = "day28coded", y = "LogLDH",
color = "day28coded", palette = c("#00AFBB", "#E7B800", "cyan2"),
ylab = "LogLDH", xlab = "day28coded")
ggboxplot(CART15data, x = "Max Grade NTX", y = "LogLDH",
color = "Max Grade NTX", palette = c("#00AFBB", "#E7B800", "cyan2", "pink"),
ylab = "LogLDH", xlab = "Max Grade NTX")
ggboxplot(CART15data, x = "Max Grade CRS", y = "LogLDH",
color = "Max Grade CRS", palette = c("#00AFBB", "#E7B800", "cyan2", "pink"),
ylab = "LogLDH", xlab = "Max Grade CRS")
##Cytokines
spBCA1<- cor.test(CART15data$LogBCA1, CART15data$LogLDH, method =c('pearson'))
spFrac<- cor.test(CART15data$LogFractalkine, CART15data$LogLDH, method =c('pearson'))
spGCSF<- cor.test(CART15data$LogGCSF, CART15data$LogLDH, method =c('pearson'))
spI309<- cor.test(CART15data$LogI309, CART15data$LogLDH, method =c('pearson'))
spIFNg<- cor.test(CART15data$LogIFNg, CART15data$LogLDH, method =c('pearson'))
spIL6<- cor.test(CART15data$LogIL6, CART15data$LogLDH, method =c('pearson'))
spIL8<- cor.test(CART15data$LogIL8, CART15data$LogLDH, method =c('pearson'))
spIP10<- cor.test(CART15data$LogIP10, CART15data$LogLDH, method =c('pearson'))
spMCP2<- cor.test(CART15data$LogMCP2, CART15data$LogLDH, method =c('pearson'))
spTNFa<- cor.test(CART15data$LogTNFa, CART15data$LogLDH, method =c('pearson'))
data.frame(spBCA1$p.value, spFrac$p.value, spGCSF$p.value, spI309$p.value, spIFNg$p.value, spIL6$p.value, spIL8$p.value, spIP10$p.value, spMCP2$p.value, spTNFa$p.value)#KYN Metabolites
pcTrp<- cor.test(CARTdata$LogTRP, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcKyn <- cor.test(CARTdata$LogKYN, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcKA <- cor.test(CARTdata$LogKA, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcHK <- cor.test(CARTdata$LogHK, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcHAA <- cor.test(CARTdata$LogHAA2, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcQA <- cor.test(CARTdata$LogQA, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
data.frame(pcTrp$p.value, pcKyn$p.value, pcKA$p.value, pcHK$p.value, pcHAA$p.value, pcQA$p.value)#PROs
pcDep<- cor.test(CARTdata$Dep, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcAnx <- cor.test(CARTdata$Anx, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcPSQI <- cor.test(CARTdata$PSQI, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcBPII <- cor.test(CARTdata$BPII, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcBPIF <- cor.test(CARTdata$BPIF, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcFSII <- cor.test(CARTdata$FSII, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcFSID <- cor.test(CARTdata$FSID, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
pcFSIF <- cor.test(CARTdata$FSIF, CARTdata$LogLDH * CARTdata$Time, method =c('pearson'))
data.frame(pcDep$p.value, pcAnx$p.value, pcPSQI$p.value, pcBPII$p.value, pcBPIF$p.value, pcFSII$p.value, pcFSID$p.value, pcFSIF$p.value)LDH as Covariant
#Demographics
lmAge <-lm(CART15data$Age ~ CART15data$IncomeDi + CART15data$LogLDH)
lmEd <-glm(CART15data$EducationLevelCoded ~ CART15data$IncomeDi + CART15data$LogLDH)
lmLines <-lm(CART15data$lines ~ CART15data$IncomeDi + CART15data$LogLDH)
lmCR <-glm(CART15data$day28coded ~ CART15data$IncomeDi + CART15data$LogLDH)
lmCRS <-glm(CART15data$`CRS (Y=1 /N =0)` ~ CART15data$IncomeDi + CART15data$LogLDH, family=binomial)
lmCRSgrade <-glm(CART15data$`Max Grade CRS` ~ CART15data$IncomeDi + CART15data$LogLDH)
lmCRSdays <-lm(CART15data$`Day to CRS`~ CART15data$IncomeDi + CART15data$LogLDH)
lmNTX <-glm(CART15data$`NTX (yes=1, no =0)` ~ CART15data$IncomeDi + CART15data$LogLDH, family=binomial)
lmNTXgrade <-glm(CART15data$`Max Grade NTX` ~ CART15data$IncomeDi + CART15data$LogLDH)
lmNTXdays <-lm(CART15data$`Day to NTX` ~ CART15data$IncomeDi + CART15data$LogLDH)
#Cytokines
lmBCA1<- lm(CART15data$LogBCA1 ~ CART15data$IncomeDi + CART15data$LogLDH)
lmFrac<- lm(CART15data$LogFractalkine ~ CART15data$IncomeDi + CART15data$LogLDH)
lmGCSF<- lm(CART15data$LogGCSF ~ CART15data$IncomeDi + CART15data$LogLDH)
lmI309<- lm(CART15data$LogI309 ~ CART15data$IncomeDi + CART15data$LogLDH)
lmIFNg<- lm(CART15data$LogIFNg ~ CART15data$IncomeDi + CART15data$LogLDH)
lmIL6<- lm(CART15data$LogIL6 ~ CART15data$IncomeDi + CART15data$LogLDH)
lmIL8<- lm(CART15data$LogIL8 ~ CART15data$IncomeDi + CART15data$LogLDH)
lmIP10<- lm(CART15data$LogIP10 ~ CART15data$IncomeDi + CART15data$LogLDH)
lmMCP2<- lm(CART15data$LogMCP2 ~ CART15data$IncomeDi + CART15data$LogLDH)
lmTNFa<- lm(CART15data$LogTNFa ~ CART15data$IncomeDi + CART15data$LogLDH)
#KYN Metabolites
lmrTRP <- lmer(LogTRP~IncomeDiCode*TimeCode + CARTdata$LogLDH + (1|ID), data=CARTdata)
emTRP<- emmeans(lmrTRP, pairwise~IncomeDiCode|TimeCode)
lmrKYN <- lmer(LogKYN~IncomeDiCode*TimeCode+ CARTdata$LogLDH +(1|ID), data=CARTdata)
emKYN<- emmeans(lmrKYN, pairwise~IncomeDiCode|TimeCode)
lmrKA <- lmer(LogKA~IncomeDiCode*TimeCode+ CARTdata$LogLDH +(1|ID), data=CARTdata)
emKA<- emmeans(lmrKA, pairwise~IncomeDiCode|TimeCode)
lmrHK <- lmer(LogHK~IncomeDiCode*TimeCode+ CARTdata$LogLDH + (1|ID), data=CARTdata)
emHK<- emmeans(lmrHK, pairwise~IncomeDiCode|TimeCode)
lmrHAA <- lmer(LogHAA2~IncomeDiCode*TimeCode+CARTdata$LogLDH + (1|ID), data=CARTdata)
emHAA<- emmeans(lmrHAA, pairwise~IncomeDiCode|TimeCode)
lmrQA <- lmer(LogQA~IncomeDiCode*TimeCode+CARTdata$LogLDH + (1|ID), data=CARTdata)
emQA<- emmeans(lmrQA, pairwise~IncomeDiCode|TimeCode)
#PROs
lmrDep <- lmer(Dep~IncomeDiCode*TimeCode+ CARTdata$LogLDH+(1|ID), data=CARTdata)
emDep<- emmeans(lmrDep, pairwise~IncomeDiCode|TimeCode)
lmrAnx <- lmer(Anx~IncomeDiCode*TimeCode+ CARTdata$LogLDH+(1|ID), data=CARTdata)
emAnx<- emmeans(lmrAnx, pairwise~IncomeDiCode|TimeCode)
lmrPSQI <- lmer(PSQI~IncomeDiCode*TimeCode+ CARTdata$LogLDH+(1|ID), data=CARTdata)
emPSQI<- emmeans(lmrPSQI, pairwise~IncomeDiCode|TimeCode)
lmrFSII <- lmer(FSII~IncomeDiCode*TimeCode+ CARTdata$LogLDH+(1|ID), data=CARTdata)
emFSII<- emmeans(lmrFSII, pairwise~IncomeDiCode|TimeCode)
lmrFSIF <- lmer(FSIF~IncomeDiCode*TimeCode+ CARTdata$LogLDH+(1|ID), data=CARTdata)
emFSIF<- emmeans(lmrFSIF, pairwise~IncomeDiCode|TimeCode)
lmrFSID <- lmer(FSID~IncomeDiCode*TimeCode+ CARTdata$LogLDH+(1|ID), data=CARTdata)
emFSID<- emmeans(lmrFSID, pairwise~IncomeDiCode|TimeCode)
lmrBPII <- lmer(BPII~IncomeDiCode*TimeCode+ CARTdata$LogLDH+(1|ID), data=CARTdata)
emBPII<- emmeans(lmrBPII, pairwise~IncomeDiCode|TimeCode)
lmrBPIF <- lmer(BPIF~IncomeDiCode*TimeCode+ CARTdata$LogLDH+(1|ID), data=CARTdata)
emBPIF<- emmeans(lmrBPIF, pairwise~IncomeDiCode|TimeCode)
#Data Output
summary(lmAge)##
## Call:
## lm(formula = CART15data$Age ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -16.889 -9.670 3.474 7.732 13.440
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 52.489 31.056 1.690 0.117
## CART15data$IncomeDi -5.016 6.877 -0.729 0.480
## CART15data$LogLDH 5.393 9.972 0.541 0.599
##
## Residual standard error: 11.43 on 12 degrees of freedom
## Multiple R-squared: 0.1213, Adjusted R-squared: -0.02519
## F-statistic: 0.828 on 2 and 12 DF, p-value: 0.4604summary(lmEd)##
## Call:
## glm(formula = CART15data$EducationLevelCoded ~ CART15data$IncomeDi +
## CART15data$LogLDH)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.6670 -0.6298 0.1805 0.6136 1.5312
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -7.0851 2.7365 -2.589 0.02370 *
## CART15data$IncomeDi 1.9810 0.6060 3.269 0.00671 **
## CART15data$LogLDH 2.9454 0.8787 3.352 0.00576 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 1.015245)
##
## Null deviance: 26.933 on 14 degrees of freedom
## Residual deviance: 12.183 on 12 degrees of freedom
## AIC: 47.448
##
## Number of Fisher Scoring iterations: 2summary(lmLines)##
## Call:
## lm(formula = CART15data$lines ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2.8320 -1.8963 -0.5086 1.9488 4.2037
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -4.649 6.683 -0.696 0.500
## CART15data$IncomeDi 1.048 1.480 0.708 0.493
## CART15data$LogLDH 3.135 2.146 1.461 0.170
##
## Residual standard error: 2.461 on 12 degrees of freedom
## Multiple R-squared: 0.1511, Adjusted R-squared: 0.00958
## F-statistic: 1.068 on 2 and 12 DF, p-value: 0.3743summary(lmCR)##
## Call:
## glm(formula = CART15data$day28coded ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.79413 -0.27857 -0.19429 -0.03778 1.57848
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -1.1120 1.9706 -0.564 0.583
## CART15data$IncomeDi 0.2148 0.4364 0.492 0.631
## CART15data$LogLDH 0.8658 0.6328 1.368 0.196
##
## (Dispersion parameter for gaussian family taken to be 0.5264882)
##
## Null deviance: 7.3333 on 14 degrees of freedom
## Residual deviance: 6.3179 on 12 degrees of freedom
## AIC: 37.598
##
## Number of Fisher Scoring iterations: 2summary(lmCRS)##
## Call:
## glm(formula = CART15data$`CRS (Y=1 /N =0)` ~ CART15data$IncomeDi +
## CART15data$LogLDH, family = binomial)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.52038 0.00000 0.00003 0.21798 1.19848
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 12.28 9041.64 0.001 0.999
## CART15data$IncomeDi -19.92 4520.80 -0.004 0.996
## CART15data$LogLDH 12.29 14.35 0.856 0.392
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 15.0121 on 14 degrees of freedom
## Residual deviance: 8.5456 on 12 degrees of freedom
## AIC: 14.546
##
## Number of Fisher Scoring iterations: 19summary(lmCRSgrade)##
## Call:
## glm(formula = CART15data$`Max Grade CRS` ~ CART15data$IncomeDi +
## CART15data$LogLDH)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.13688 -0.63402 -0.02038 0.29446 2.16806
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.1238 2.5427 0.442 0.666
## CART15data$IncomeDi -0.8395 0.5631 -1.491 0.162
## CART15data$LogLDH 0.5585 0.8165 0.684 0.507
##
## (Dispersion parameter for gaussian family taken to be 0.8765312)
##
## Null deviance: 14.933 on 14 degrees of freedom
## Residual deviance: 10.518 on 12 degrees of freedom
## AIC: 45.244
##
## Number of Fisher Scoring iterations: 2summary(lmCRSdays)##
## Call:
## lm(formula = CART15data$`Day to CRS` ~ CART15data$IncomeDi +
## CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -14.940 -4.420 -1.278 3.615 12.220
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -3.570 22.983 -0.155 0.8791
## CART15data$IncomeDi 12.401 5.089 2.437 0.0313 *
## CART15data$LogLDH -2.365 7.380 -0.320 0.7541
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 8.463 on 12 degrees of freedom
## Multiple R-squared: 0.4348, Adjusted R-squared: 0.3406
## F-statistic: 4.615 on 2 and 12 DF, p-value: 0.03261summary(lmNTX)##
## Call:
## glm(formula = CART15data$`NTX (yes=1, no =0)` ~ CART15data$IncomeDi +
## CART15data$LogLDH, family = binomial)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.3455 -0.6563 -0.5135 0.3986 2.0495
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -10.1685 8.2515 -1.232 0.218
## CART15data$IncomeDi -0.4987 1.5918 -0.313 0.754
## CART15data$LogLDH 4.1090 2.8292 1.452 0.146
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 19.095 on 14 degrees of freedom
## Residual deviance: 13.465 on 12 degrees of freedom
## AIC: 19.465
##
## Number of Fisher Scoring iterations: 4summary(lmNTXgrade)##
## Call:
## glm(formula = CART15data$`Max Grade NTX` ~ CART15data$IncomeDi +
## CART15data$LogLDH)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.56414 -0.45774 -0.08543 0.20704 2.20487
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -5.9140 2.4728 -2.392 0.0340 *
## CART15data$IncomeDi -0.2262 0.5476 -0.413 0.6868
## CART15data$LogLDH 2.8639 0.7941 3.607 0.0036 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.829032)
##
## Null deviance: 26.4000 on 14 degrees of freedom
## Residual deviance: 9.9484 on 12 degrees of freedom
## AIC: 44.409
##
## Number of Fisher Scoring iterations: 2summary(lmNTXdays)##
## Call:
## lm(formula = CART15data$`Day to NTX` ~ CART15data$IncomeDi +
## CART15data$LogLDH)
##
## Residuals:
## 2 9 11 13 15
## 1.267e+00 -1.104e+00 3.526e+00 5.551e-16 -3.690e+00
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 25.078 18.298 1.370 0.304
## CART15data$IncomeDi -2.189 5.303 -0.413 0.720
## CART15data$LogLDH -6.568 4.978 -1.319 0.318
##
## Residual standard error: 3.799 on 2 degrees of freedom
## (10 observations deleted due to missingness)
## Multiple R-squared: 0.4953, Adjusted R-squared: -0.009494
## F-statistic: 0.9812 on 2 and 2 DF, p-value: 0.5047summary(lmBCA1)##
## Call:
## lm(formula = CART15data$LogBCA1 ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.61326 -0.37604 -0.03261 0.20323 0.86373
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.3117 1.4545 0.902 0.385
## CART15data$IncomeDi -0.2922 0.3221 -0.907 0.382
## CART15data$LogLDH 0.6874 0.4671 1.472 0.167
##
## Residual standard error: 0.5356 on 12 degrees of freedom
## Multiple R-squared: 0.3286, Adjusted R-squared: 0.2167
## F-statistic: 2.936 on 2 and 12 DF, p-value: 0.09162summary(lmFrac)##
## Call:
## lm(formula = CART15data$LogFractalkine ~ CART15data$IncomeDi +
## CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.9618 -0.4429 -0.2082 0.4501 1.3175
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.18511 1.90406 0.622 0.545
## CART15data$IncomeDi -0.03405 0.42164 -0.081 0.937
## CART15data$LogLDH 0.54646 0.61141 0.894 0.389
##
## Residual standard error: 0.7011 on 12 degrees of freedom
## Multiple R-squared: 0.09, Adjusted R-squared: -0.06167
## F-statistic: 0.5934 on 2 and 12 DF, p-value: 0.5679summary(lmGCSF)##
## Call:
## lm(formula = CART15data$LogGCSF ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.8997 -0.5552 -0.0602 0.4961 0.8597
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.5409 1.6952 0.319 0.7552
## CART15data$IncomeDi -0.4572 0.3754 -1.218 0.2466
## CART15data$LogLDH 0.9793 0.5444 1.799 0.0972 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.6242 on 12 degrees of freedom
## Multiple R-squared: 0.4389, Adjusted R-squared: 0.3454
## F-statistic: 4.694 on 2 and 12 DF, p-value: 0.0312summary(lmI309)##
## Call:
## lm(formula = CART15data$LogI309 ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.73475 -0.20145 -0.06731 0.23076 0.52402
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.4551 0.9929 0.458 0.6548
## CART15data$IncomeDi -0.2926 0.2199 -1.331 0.2079
## CART15data$LogLDH 0.7293 0.3188 2.287 0.0411 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3656 on 12 degrees of freedom
## Multiple R-squared: 0.5321, Adjusted R-squared: 0.4541
## F-statistic: 6.823 on 2 and 12 DF, p-value: 0.0105summary(lmIFNg)##
## Call:
## lm(formula = CART15data$LogIFNg ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.33371 -0.55690 0.09247 0.36874 1.57659
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.8564 2.4247 0.353 0.730
## CART15data$IncomeDi -0.3444 0.5369 -0.641 0.533
## CART15data$LogLDH 0.7586 0.7786 0.974 0.349
##
## Residual standard error: 0.8928 on 12 degrees of freedom
## Multiple R-squared: 0.1835, Adjusted R-squared: 0.04742
## F-statistic: 1.348 on 2 and 12 DF, p-value: 0.2963summary(lmIL6)##
## Call:
## lm(formula = CART15data$LogIL6 ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.08332 -0.47628 0.07498 0.37896 1.40892
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.0843 1.8755 0.578 0.574
## CART15data$IncomeDi -0.4670 0.4153 -1.125 0.283
## CART15data$LogLDH 0.6659 0.6022 1.106 0.291
##
## Residual standard error: 0.6905 on 12 degrees of freedom
## Multiple R-squared: 0.297, Adjusted R-squared: 0.1798
## F-statistic: 2.535 on 2 and 12 DF, p-value: 0.1207summary(lmIL8)##
## Call:
## lm(formula = CART15data$LogIL8 ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.48514 -0.31753 -0.02474 0.20745 0.89244
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.2113 1.1540 1.050 0.315
## CART15data$IncomeDi -0.3912 0.2555 -1.531 0.152
## CART15data$LogLDH 0.5395 0.3705 1.456 0.171
##
## Residual standard error: 0.4249 on 12 degrees of freedom
## Multiple R-squared: 0.4311, Adjusted R-squared: 0.3363
## F-statistic: 4.546 on 2 and 12 DF, p-value: 0.03391summary(lmIP10)##
## Call:
## lm(formula = CART15data$LogIP10 ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.47377 -0.13282 0.02447 0.13102 0.57869
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.97844 0.88727 4.484 0.000747 ***
## CART15data$IncomeDi -0.50612 0.19648 -2.576 0.024277 *
## CART15data$LogLDH 0.06551 0.28491 0.230 0.822028
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3267 on 12 degrees of freedom
## Multiple R-squared: 0.4507, Adjusted R-squared: 0.3592
## F-statistic: 4.923 on 2 and 12 DF, p-value: 0.02747summary(lmMCP2)##
## Call:
## lm(formula = CART15data$LogMCP2 ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.41567 -0.20554 -0.06313 0.18158 0.76058
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.1369 0.9475 2.255 0.0436 *
## CART15data$IncomeDi -0.3527 0.2098 -1.681 0.1186
## CART15data$LogLDH 0.1957 0.3043 0.643 0.5322
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3489 on 12 degrees of freedom
## Multiple R-squared: 0.3282, Adjusted R-squared: 0.2162
## F-statistic: 2.931 on 2 and 12 DF, p-value: 0.09195summary(lmTNFa)##
## Call:
## lm(formula = CART15data$LogTNFa ~ CART15data$IncomeDi + CART15data$LogLDH)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.40389 -0.20705 -0.09129 0.16745 0.86054
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.8191 0.9485 0.863 0.4048
## CART15data$IncomeDi -0.3497 0.2100 -1.665 0.1218
## CART15data$LogLDH 0.8361 0.3046 2.745 0.0178 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3493 on 12 degrees of freedom
## Multiple R-squared: 0.6274, Adjusted R-squared: 0.5653
## F-statistic: 10.1 on 2 and 12 DF, p-value: 0.002677emTRP$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low 0.1500 0.131 36.6 1.142 0.2610
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low 0.1387 0.144 40.6 0.963 0.3411
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low 0.0109 0.137 38.1 0.079 0.9374
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low 0.0928 0.138 38.1 0.673 0.5052
##
## Degrees-of-freedom method: kenward-rogeremKYN$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -0.1719 0.140 31.9 -1.226 0.2293
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low -0.0909 0.152 36.6 -0.598 0.5538
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -0.0942 0.146 33.7 -0.645 0.5232
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -0.0750 0.147 33.8 -0.511 0.6123
##
## Degrees-of-freedom method: kenward-rogeremKA$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low 0.18246 0.213 29.3 0.858 0.3980
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low 0.06580 0.229 34.1 0.287 0.7760
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low 0.00732 0.221 31.3 0.033 0.9738
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low 0.21614 0.222 31.4 0.975 0.3372
##
## Degrees-of-freedom method: kenward-rogeremHK$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -0.1324 0.281 17.5 -0.472 0.6428
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low -0.0958 0.291 19.7 -0.329 0.7453
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -0.1697 0.286 18.6 -0.594 0.5600
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low 0.1436 0.286 18.7 0.501 0.6219
##
## Degrees-of-freedom method: kenward-rogeremHAA$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -0.348 0.439 20.0 -0.792 0.4378
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low -0.456 0.460 23.1 -0.992 0.3315
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -0.668 0.450 21.4 -1.485 0.1522
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -0.183 0.451 21.6 -0.407 0.6884
##
## Degrees-of-freedom method: kenward-rogeremQA$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -0.633 0.218 12.5 -2.897 0.0129
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low -0.747 0.284 13.0 -2.627 0.0209
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -0.548 0.255 12.7 -2.146 0.0519
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -0.436 0.284 12.7 -1.537 0.1487
##
## Degrees-of-freedom method: kenward-rogeremDep$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -1.72 6.21 23.9 -0.277 0.7841
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low 7.81 7.85 34.7 0.995 0.3265
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -4.50 6.40 25.6 -0.702 0.4890
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -2.52 6.70 28.0 -0.376 0.7096
##
## Degrees-of-freedom method: kenward-rogeremAnx$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -1.4385 3.52 18.1 -0.408 0.6878
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low -0.8964 4.14 27.6 -0.217 0.8301
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low 0.0964 3.59 19.2 0.027 0.9789
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -2.5600 3.70 21.0 -0.692 0.4965
##
## Degrees-of-freedom method: kenward-rogeremPSQI$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -3.12 1.95 26.3 -1.603 0.1208
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low -2.48 2.60 34.7 -0.956 0.3455
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -2.18 2.11 29.1 -1.037 0.3083
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -4.51 2.21 31.0 -2.039 0.0500
##
## Degrees-of-freedom method: kenward-rogeremFSII$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -1.555 1.12 23.0 -1.393 0.1769
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low 0.948 1.40 34.0 0.678 0.5023
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -0.659 1.15 24.6 -0.574 0.5711
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -1.055 1.20 27.0 -0.880 0.3868
##
## Degrees-of-freedom method: kenward-rogeremFSIF$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -1.02 1.16 28.4 -0.883 0.3844
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low 1.07 1.51 37.1 0.708 0.4836
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -1.90 1.20 30.0 -1.589 0.1226
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -0.65 1.26 32.3 -0.514 0.6108
##
## Degrees-of-freedom method: kenward-rogeremFSID$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -0.569 2.88 22.4 -0.197 0.8452
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low 7.442 3.58 33.4 2.077 0.0456
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -1.105 2.96 23.9 -0.373 0.7123
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low 0.823 3.09 26.3 0.267 0.7919
##
## Degrees-of-freedom method: kenward-rogeremBPII$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -2.548 1.09 21.8 -2.340 0.0289
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low -0.337 1.35 32.8 -0.251 0.8037
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -2.826 1.12 23.3 -2.527 0.0188
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -3.676 1.16 25.6 -3.158 0.0040
##
## Degrees-of-freedom method: kenward-rogeremBPIF$contrasts## TimeCode = Baseline:
## contrast estimate SE df t.ratio p.value
## high - low -2.97 1.34 20.4 -2.214 0.0384
##
## TimeCode = D14:
## contrast estimate SE df t.ratio p.value
## high - low -1.38 1.63 31.1 -0.849 0.4023
##
## TimeCode = D28:
## contrast estimate SE df t.ratio p.value
## high - low -2.69 1.38 21.8 -1.955 0.0635
##
## TimeCode = D90:
## contrast estimate SE df t.ratio p.value
## high - low -2.23 1.43 23.9 -1.561 0.1317
##
## Degrees-of-freedom method: kenward-rogerThe basic approach is regression adjustment: -continuous demographics: t-test/ANOVA is equivalent to lm(Age ~ IncomeDiCode, data=CAR15data), so we do lm(Age ~ IncomeDiCode+ LogLDH, data=CAR15data)
bi-variate demographics: chi-square test is equivalent to logistic regression glm(Sex ~ SES, data=CAR15data, family=binomial), so we do glm(Sex ~ IncomeDiCode+ LogLDH, data=CAR15data, family=binomial).
for repeated measures you did something like lmer(Outcome ~ SESTime + (1|subject), data=CARTdata), so here you would do lmer(Outcome ~ IncomeDiCodeTime + LogLDH + (1|subject), data=CARTdata).