Evaluation des conditions de validité:Normalité pour les données IgG 2 deficit Vs Normaux + Excès

la normalité va nous permettre de choisir le type de test pour comparer les données Un test de shapiro pour valider la normalité si la p-value est<0.05 la normalité n’est pas accepté on fait donc un test de Wilcoxon Si la normalité est respecté on fait le test de Student

Age à l’inclusion

shapiro.test(Ig$age_inclusion[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$age_inclusion[Ig$IgG2_Exces == "Oui"]
## W = 0.95778, p-value = 0.2093

Wilcoxon

shapiro.test(Ig$age_inclusion[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$age_inclusion[Ig$IgG2_Exces == "Non"]
## W = 0.97372, p-value = 0.000349

student

IMC

shapiro.test(Ig$IMC[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$IMC[Ig$IgG2_Exces == "Oui"]
## W = 0.67554, p-value = 9.712e-06

la normalité est acceptée: T-test de Student

shapiro.test(Ig$IMC[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$IMC[Ig$IgG2_Exces == "Non"]
## W = 0.95126, p-value = 5.299e-05

student

Nombre d’exacerbation sans biotherapie

shapiro.test(Ig$Nb_Exa_an.sans_biotherapie[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Nb_Exa_an.sans_biotherapie[Ig$IgG2_Exces == "Oui"]
## W = 0.86295, p-value = 0.002102

student

shapiro.test(Ig$Nb_Exa_an.sans_biotherapie[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Nb_Exa_an.sans_biotherapie[Ig$IgG2_Exces == "Non"]
## W = 0.37663, p-value < 2.2e-16

Student

Dose de corticoides systemiques

shapiro.test(Ig$Corticoïde.systémique..dose.[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Corticoïde.systémique..dose.[Ig$IgG2_Exces == "Oui"]
## W = 0.43556, p-value = 1.763e-09

Student

shapiro.test(Ig$Corticoïde.systémique..dose.[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Corticoïde.systémique..dose.[Ig$IgG2_Exces == "Non"]
## W = 0.52583, p-value < 2.2e-16

Student

CSI

shapiro.test(Ig$CSI_µg[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$CSI_µg[Ig$IgG2_Exces == "Oui"]
## W = 0.85205, p-value = 0.0005605

la normalité est acceptée : T test de student

shapiro.test(Ig$CSI_µg[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$CSI_µg[Ig$IgG2_Exces == "Non"]
## W = 0.87501, p-value = 2.376e-12

Student

Montélukast

shapiro.test(Ig$Montélukast[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Montélukast[Ig$IgG2_Exces == "Oui"]
## W = 0.63269, p-value = 5.288e-08

Student

shapiro.test(Ig$Montélukast[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Montélukast[Ig$IgG2_Exces == "Non"]
## W = 0.60744, p-value < 2.2e-16

Student

Réponse Biotherapie

shapiro.test(Ig$Rep_biotherapie_GETE[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Rep_biotherapie_GETE[Ig$IgG2_Exces == "Oui"]
## W = 0.58141, p-value = 4.361e-08

la normalité est acceptée : t-test de Student

shapiro.test(Ig$Rep_biotherapie_GETE[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Rep_biotherapie_GETE[Ig$IgG2_Exces == "Non"]
## W = 0.65403, p-value < 2.2e-16

Student

VEMS pre B2 (litre)

shapiro.test(Ig$VEMS_pre_B2_L[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$VEMS_pre_B2_L[Ig$IgG2_Exces == "Oui"]
## W = 0.93689, p-value = 0.09211

Wilcoxon

shapiro.test(Ig$VEMS_pre_B2_L[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$VEMS_pre_B2_L[Ig$IgG2_Exces == "Non"]
## W = 0.97659, p-value = 0.002462

Student

VEMS pre B2 (%)

shapiro.test(Ig$VEMS_PreB2_Pct[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$VEMS_PreB2_Pct[Ig$IgG2_Exces == "Oui"]
## W = 0.92247, p-value = 0.03524

la normalité est acceptée : t-test de Student

shapiro.test(Ig$VEMS_PreB2_Pct[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$VEMS_PreB2_Pct[Ig$IgG2_Exces == "Non"]
## W = 0.97736, p-value = 0.002726

Student

Tiffe

shapiro.test(Ig$Tiffenau[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Tiffenau[Ig$IgG2_Exces == "Oui"]
## W = 0.95848, p-value = 0.3208

Wilcoxon

shapiro.test(Ig$Tiffenau[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$Tiffenau[Ig$IgG2_Exces == "Non"]
## W = 0.98974, p-value = 0.1766

Wilcoxon

FeNO

shapiro.test(Ig$FeNo[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$FeNo[Ig$IgG2_Exces == "Non"]
## W = 0.80675, p-value = 1.999e-11

Student

shapiro.test(Ig$FeNo[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$FeNo[Ig$IgG2_Exces == "Oui"]
## W = 0.90184, p-value = 0.1015

Wilcoxon

PNE G/L

shapiro.test(Ig$PNE_G_L[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$PNE_G_L[Ig$IgG2_Exces == "Oui"]
## W = 0.77671, p-value = 1.969e-05

Student

shapiro.test(Ig$PNE_G_L[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$PNE_G_L[Ig$IgG2_Exces == "Non"]
## W = 0.67157, p-value < 2.2e-16

Student

IgGE total

shapiro.test(Ig$IgE_Total[Ig$IgG2_Exces =="Oui"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$IgE_Total[Ig$IgG2_Exces == "Oui"]
## W = 0.88108, p-value = 0.003555

Student

shapiro.test(Ig$IgE_Total[Ig$IgG2_Exces =="Non"])
## 
##  Shapiro-Wilk normality test
## 
## data:  Ig$IgE_Total[Ig$IgG2_Exces == "Non"]
## W = 0.31944, p-value < 2.2e-16

Student

Comparaison des données IgG deficit Vs Normaux + Excès

Age

wilcox.test(Ig$age_inclusion~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  Ig$age_inclusion by Ig$IgG2_Exces
## W = 3734.5, p-value = 0.8571
## alternative hypothesis: true location shift is not equal to 0

L’hypothèse nulle d’égalité des moyennes est acceptée car la p-value est > 0.05. Le test ne met pas en évidence une différence significative

IMC

t.test(Ig$IMC~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$IMC by Ig$IgG2_Exces
## t = -1.9671, df = 166, p-value = 0.05084
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -6.22953783  0.01148054
## sample estimates:
## mean in group Non mean in group Oui 
##          26.16370          29.27273

L’hypothèse nulle d’égalité des moyennes n’est pas rejetée car la p-value est =0.05.

Nombre d’exacerbation sans biotherapie

t.test(Ig$Nb_Exa_an.sans_biotherapie~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$Nb_Exa_an.sans_biotherapie by Ig$IgG2_Exces
## t = -0.76459, df = 207, p-value = 0.4454
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -4.759652  2.099497
## sample estimates:
## mean in group Non mean in group Oui 
##          4.373626          5.703704

Le test ne met pas en évidence une différence significative

Dose de corticoides systemiques

t.test(Ig$Corticoïde.systémique..dose.~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$Corticoïde.systémique..dose. by Ig$IgG2_Exces
## t = 0.45784, df = 229, p-value = 0.6475
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -3.974139  6.380080
## sample estimates:
## mean in group Non mean in group Oui 
##           6.20297           5.00000

Le test ne met pas en évidence une différence significative

CSI

t.test(Ig$CSI_µg~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$CSI_µg by Ig$IgG2_Exces
## t = -0.82374, df = 245, p-value = 0.4109
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -679.5425  278.7707
## sample estimates:
## mean in group Non mean in group Oui 
##          1580.259          1780.645

Le test ne met pas en évidence une différence significative

Montélukast

t.test(Ig$Montélukast~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$Montélukast by Ig$IgG2_Exces
## t = -0.90627, df = 254, p-value = 0.3657
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.25643178  0.09479956
## sample estimates:
## mean in group Non mean in group Oui 
##         0.3603604         0.4411765

Le test ne met pas en évidence une différence significative

Réponse Biotherapie

t.test(Ig$Rep_biotherapie_GETE~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$Rep_biotherapie_GETE by Ig$IgG2_Exces
## t = 0.57574, df = 242, p-value = 0.5653
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.4375048  0.7988755
## sample estimates:
## mean in group Non mean in group Oui 
##         1.0140187         0.8333333

Le test ne met pas en évidence une différence significative

VEMS pre B2 (litre)

wilcox.test(Ig$VEMS_pre_B2_L~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  Ig$VEMS_pre_B2_L by Ig$IgG2_Exces
## W = 2949.5, p-value = 0.4633
## alternative hypothesis: true location shift is not equal to 0

Le test ne met pas en évidence une différence significative

VEMS pre B2 (%)

t.test(Ig$VEMS_PreB2_Pct~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$VEMS_PreB2_Pct by Ig$IgG2_Exces
## t = 1.1171, df = 225, p-value = 0.2651
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -4.098335 14.826652
## sample estimates:
## mean in group Non mean in group Oui 
##          73.78485          68.42069

Le test ne met pas en évidence une différence significative

Tiffe

wilcox.test(Ig$Tiffenau~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  Ig$Tiffenau by Ig$IgG2_Exces
## W = 3030, p-value = 0.3481
## alternative hypothesis: true location shift is not equal to 0

Le test ne met pas en évidence une différence significative

FeNO

malgre les données insuffisantes, le test de wilcoxon a été fait ( attention à ne pas prendre en consideration)

wilcox.test(Ig$FeNo~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  Ig$FeNo by Ig$IgG2_Exces
## W = 1148.5, p-value = 0.1228
## alternative hypothesis: true location shift is not equal to 0

Le test ne met pas en évidence une différence significative

PNE_Gl

t.test(Ig$PNE_G_L~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$PNE_G_L by Ig$IgG2_Exces
## t = -0.064516, df = 233, p-value = 0.9486
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.1760932  0.1649262
## sample estimates:
## mean in group Non mean in group Oui 
##         0.3460294         0.3516129

Le test ne met pas en évidence une différence significative

IgGE total

t.test(Ig$IgE_Total~Ig$IgG2_Exces, var.equal=TRUE)
## 
##  Two Sample t-test
## 
## data:  Ig$IgE_Total by Ig$IgG2_Exces
## t = 0.46644, df = 179, p-value = 0.6415
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -312.9386  506.6763
## sample estimates:
## mean in group Non mean in group Oui 
##          425.9392          329.0703

la p-value est dans la limite de la significativité

Test de comparaison des données qualitatives: IgG deficit Vs normaux+exces

comp_qual <- read.csv("C:/Users/mallah.s/Desktop/StatsTheses/Mauro anthony/comp_qual.csv", sep=";", stringsAsFactors=TRUE)

Pour le test de liaison entre deux variables qualitative: verification de la normalité -> selon la validation de la normalité :test de Chi2 ou test de fisher

le test du χ2 d’indépendance sert à étudier la liaison entre deux caractères qualitatifs XetY, lorsque les conditions ne sont pas remplies, il existe des corrections,dans notre cas je vais utiliser le tests exacts de Fisher

Sexe

xtabs(~Sexe+IgG2_Exces, data=comp_qual)
##     IgG2_Exces
## Sexe Non Oui
##    F 134  20
##    M  90  14
chisq.test(comp_qual$Sexe,comp_qual$IgG2_Exces)$expected
##               comp_qual$IgG2_Exces
## comp_qual$Sexe       Non      Oui
##              F 133.70543 20.29457
##              M  90.29457 13.70543
chisq.test(comp_qual$Sexe,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$Sexe and comp_qual$IgG2_Exces
## X-squared = 0.012217, df = 1, p-value = 0.912
fisher.test(comp_qual$Sexe,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$Sexe and comp_qual$IgG2_Exces
## p-value = 1
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.4611336 2.2985722
## sample estimates:
## odds ratio 
##   1.042086

les variables Sexe et deficit d’IgG sont independantes

Montélukast

xtabs(~Montélukast+IgG2_Exces, data=comp_qual)
##            IgG2_Exces
## Montélukast Non Oui
##         Non 142  19
##         Oui  80  15
chisq.test(comp_qual$Montélukast,comp_qual$IgG2_Exces)$expected
##                      comp_qual$IgG2_Exces
## comp_qual$Montélukast       Non      Oui
##                   Non 139.61719 21.38281
##                   Oui  82.38281 12.61719
chisq.test(comp_qual$Montélukast,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$Montélukast and comp_qual$IgG2_Exces
## X-squared = 0.82512, df = 1, p-value = 0.3637
fisher.test(comp_qual$Montélukast,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$Montélukast and comp_qual$IgG2_Exces
## p-value = 0.4461
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.6245264 3.0886282
## sample estimates:
## odds ratio 
##   1.399368

Biotherapie

xtabs(~Biotherapie+IgG2_Exces, data=comp_qual)
##            IgG2_Exces
## Biotherapie Non Oui
##         Non 145  23
##         Oui  79  11
chisq.test(comp_qual$Biotherapie,comp_qual$IgG2_Exces)$expected
##                      comp_qual$IgG2_Exces
## comp_qual$Biotherapie       Non      Oui
##                   Non 145.86047 22.13953
##                   Oui  78.13953 11.86047
chisq.test(comp_qual$Biotherapie,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$Biotherapie and comp_qual$IgG2_Exces
## X-squared = 0.11042, df = 1, p-value = 0.7397
fisher.test(comp_qual$Biotherapie,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$Biotherapie and comp_qual$IgG2_Exces
## p-value = 0.8478
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.3664815 1.9921208
## sample estimates:
## odds ratio 
##  0.8782306

type de biotherapie

xtabs(~Biothérapie_type+IgG2_Exces, data=comp_qual)
##                 IgG2_Exces
## Biothérapie_type Non Oui
##     Benralizumab   9   1
##     Dupilumab      4   0
##     Mepolizumab   44   8
##     Omalixumab    22   2
chisq.test(comp_qual$Biothérapie_type,comp_qual$IgG2_Exces)$expected
## Warning in chisq.test(comp_qual$Biothérapie_type, comp_qual$IgG2_Exces): Chi-
## squared approximation may be incorrect
##                           comp_qual$IgG2_Exces
## comp_qual$Biothérapie_type       Non       Oui
##               Benralizumab  8.777778 1.2222222
##               Dupilumab     3.511111 0.4888889
##               Mepolizumab  45.644444 6.3555556
##               Omalixumab   21.066667 2.9333333
chisq.test(comp_qual$Biothérapie_type,comp_qual$IgG2_Exces, correct=FALSE)
## Warning in chisq.test(comp_qual$Biothérapie_type, comp_qual$IgG2_Exces, : Chi-
## squared approximation may be incorrect
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$Biothérapie_type and comp_qual$IgG2_Exces
## X-squared = 1.426, df = 3, p-value = 0.6994
fisher.test(comp_qual$Biothérapie_type,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$Biothérapie_type and comp_qual$IgG2_Exces
## p-value = 0.8829
## alternative hypothesis: two.sided

Tabac

xtabs(~Tabac+IgG2_Exces, data=comp_qual)
##         IgG2_Exces
## Tabac    Non Oui
##   Actif   23   2
##   Non     76  15
##   Passif   9   1
##   Sevré   75  10
chisq.test(comp_qual$Tabac,comp_qual$IgG2_Exces)$expected
## Warning in chisq.test(comp_qual$Tabac, comp_qual$IgG2_Exces): Chi-squared
## approximation may be incorrect
##                comp_qual$IgG2_Exces
## comp_qual$Tabac       Non       Oui
##          Actif  21.682464  3.317536
##          Non    78.924171 12.075829
##          Passif  8.672986  1.327014
##          Sevré  73.720379 11.279621
chisq.test(comp_qual$Tabac,comp_qual$IgG2_Exces, correct=FALSE)
## Warning in chisq.test(comp_qual$Tabac, comp_qual$IgG2_Exces, correct = FALSE):
## Chi-squared approximation may be incorrect
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$Tabac and comp_qual$IgG2_Exces
## X-squared = 1.68, df = 3, p-value = 0.6414
fisher.test(comp_qual$Tabac,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$Tabac and comp_qual$IgG2_Exces
## p-value = 0.6901
## alternative hypothesis: two.sided

Atopie

xtabs(~Atopie+IgG2_Exces, data=comp_qual)
##       IgG2_Exces
## Atopie Non Oui
##    Non 140  20
##    Oui  84  14
chisq.test(comp_qual$Atopie,comp_qual$IgG2_Exces)$expected
##                 comp_qual$IgG2_Exces
## comp_qual$Atopie       Non      Oui
##              Non 138.91473 21.08527
##              Oui  85.08527 12.91473
chisq.test(comp_qual$Atopie,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$Atopie and comp_qual$IgG2_Exces
## X-squared = 0.16938, df = 1, p-value = 0.6807
fisher.test(comp_qual$Atopie,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$Atopie and comp_qual$IgG2_Exces
## p-value = 0.7071
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.5153122 2.5759322
## sample estimates:
## odds ratio 
##   1.165954

FeNO > 20

xtabs(~FeNo_sup_20_ppb+IgG2_Exces, data=comp_qual)
##                IgG2_Exces
## FeNo_sup_20_ppb Non Oui
##             Non  49   7
##             Oui  74   8
chisq.test(comp_qual$FeNo_sup_20_ppb,comp_qual$IgG2_Exces)$expected
##                          comp_qual$IgG2_Exces
## comp_qual$FeNo_sup_20_ppb      Non      Oui
##                       Non 49.91304 6.086957
##                       Oui 73.08696 8.913043
chisq.test(comp_qual$FeNo_sup_20_ppb,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$FeNo_sup_20_ppb and comp_qual$IgG2_Exces
## X-squared = 0.2586, df = 1, p-value = 0.6111
fisher.test(comp_qual$FeNo_sup_20_ppb,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$FeNo_sup_20_ppb and comp_qual$IgG2_Exces
## p-value = 0.7815
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.2239362 2.6283393
## sample estimates:
## odds ratio 
##  0.7583267

PNN > 5

xtabs(~PNN_sup_5+IgG2_Exces, data=comp_qual)
##          IgG2_Exces
## PNN_sup_5 Non Oui
##       Non  93  13
##       Oui 113  17
chisq.test(comp_qual$PNN_sup_5,comp_qual$IgG2_Exces)$expected
##                    comp_qual$IgG2_Exces
## comp_qual$PNN_sup_5       Non      Oui
##                 Non  92.52542 13.47458
##                 Oui 113.47458 16.52542
chisq.test(comp_qual$PNN_sup_5,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$PNN_sup_5 and comp_qual$IgG2_Exces
## X-squared = 0.034762, df = 1, p-value = 0.8521
fisher.test(comp_qual$PNN_sup_5,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$PNN_sup_5 and comp_qual$IgG2_Exces
## p-value = 1
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.4646368 2.5443837
## sample estimates:
## odds ratio 
##   1.075939

Les deux variable sont independantes

PNE > 0.15

xtabs(~PNE_sup_0.15G_L+IgG2_Exces, data=comp_qual)
##                IgG2_Exces
## PNE_sup_0.15G_L Non Oui
##             Non  82  16
##             Oui 122  15
chisq.test(comp_qual$PNE_sup_0.15G_L,comp_qual$IgG2_Exces)$expected
##                          comp_qual$IgG2_Exces
## comp_qual$PNE_sup_0.15G_L       Non      Oui
##                       Non  85.07234 12.92766
##                       Oui 118.92766 18.07234
chisq.test(comp_qual$PNE_sup_0.15G_L,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$PNE_sup_0.15G_L and comp_qual$IgG2_Exces
## X-squared = 1.4428, df = 1, p-value = 0.2297
fisher.test(comp_qual$PNE_sup_0.15G_L,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$PNE_sup_0.15G_L and comp_qual$IgG2_Exces
## p-value = 0.246
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.2739374 1.4473962
## sample estimates:
## odds ratio 
##  0.6314104

PNE >0.3G/L

xtabs(~PNE_sup_0.3.G_L+IgG2_Exces, data=comp_qual)
##                IgG2_Exces
## PNE_sup_0.3.G_L Non Oui
##             Non 116  17
##             Oui  88  14
chisq.test(comp_qual$PNE_sup_0.3.G_L,comp_qual$IgG2_Exces)$expected
##                          comp_qual$IgG2_Exces
## comp_qual$PNE_sup_0.3.G_L       Non      Oui
##                       Non 115.45532 17.54468
##                       Oui  88.54468 13.45532
chisq.test(comp_qual$PNE_sup_0.3.G_L,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$PNE_sup_0.3.G_L and comp_qual$IgG2_Exces
## X-squared = 0.044879, df = 1, p-value = 0.8322
fisher.test(comp_qual$PNE_sup_0.3.G_L,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$PNE_sup_0.3.G_L and comp_qual$IgG2_Exces
## p-value = 0.848
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.4676988 2.4821256
## sample estimates:
## odds ratio 
##    1.08518

CRP >5mg/l

xtabs(~CRP_sup_5.mg_l+IgG2_Exces, data=comp_qual)
##               IgG2_Exces
## CRP_sup_5.mg_l Non Oui
##            Non  94  12
##            Oui  41   8
chisq.test(comp_qual$CRP_sup_5.mg_l,comp_qual$IgG2_Exces)$expected
##                         comp_qual$IgG2_Exces
## comp_qual$CRP_sup_5.mg_l      Non       Oui
##                      Non 92.32258 13.677419
##                      Oui 42.67742  6.322581
chisq.test(comp_qual$CRP_sup_5.mg_l,comp_qual$IgG2_Exces, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$CRP_sup_5.mg_l and comp_qual$IgG2_Exces
## X-squared = 0.74716, df = 1, p-value = 0.3874
fisher.test(comp_qual$CRP_sup_5.mg_l,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$CRP_sup_5.mg_l and comp_qual$IgG2_Exces
## p-value = 0.4421
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.5003833 4.4157556
## sample estimates:
## odds ratio 
##    1.52407

IgE totale >30 KUA/l

xtabs(~IgE_Total_sup_30_kUA_l+IgG2_Exces, data=comp_qual)
##                       IgG2_Exces
## IgE_Total_sup_30_kUA_l Non Oui
##                    Non  26   4
##                    Oui 126  25
chisq.test(comp_qual$IgE_Total_sup_30_kUA_l,comp_qual$IgG2_Exces)$expected
## Warning in chisq.test(comp_qual$IgE_Total_sup_30_kUA_l, comp_qual$IgG2_Exces):
## Chi-squared approximation may be incorrect
##      comp_qual$IgG2_Exces
##             Non      Oui
##   Non  25.19337  4.80663
##   Oui 126.80663 24.19337
chisq.test(comp_qual$IgE_Total_sup_30_kUA_l,comp_qual$IgG2_Exces, correct=FALSE)
## Warning in chisq.test(comp_qual$IgE_Total_sup_30_kUA_l, comp_qual$IgG2_Exces, :
## Chi-squared approximation may be incorrect
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$IgE_Total_sup_30_kUA_l and comp_qual$IgG2_Exces
## X-squared = 0.19322, df = 1, p-value = 0.6603
fisher.test(comp_qual$IgE_Total_sup_30_kUA_l,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$IgE_Total_sup_30_kUA_l and comp_qual$IgG2_Exces
## p-value = 0.7903
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.3952148 5.5194562
## sample estimates:
## odds ratio 
##   1.287956

Profil T2

xtabs(~Profil_T2+IgG2_Exces, data=comp_qual)
##          IgG2_Exces
## Profil_T2 Non Oui
##       Neg  73  12
##       Pos  43   2
chisq.test(comp_qual$Profil_T2,comp_qual$IgG2_Exces)$expected
## Warning in chisq.test(comp_qual$Profil_T2, comp_qual$IgG2_Exces): Chi-squared
## approximation may be incorrect
##                    comp_qual$IgG2_Exces
## comp_qual$Profil_T2      Non      Oui
##                 Neg 75.84615 9.153846
##                 Pos 40.15385 4.846154
chisq.test(comp_qual$Profil_T2,comp_qual$IgG2_Exces, correct=FALSE)
## Warning in chisq.test(comp_qual$Profil_T2, comp_qual$IgG2_Exces, correct =
## FALSE): Chi-squared approximation may be incorrect
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$Profil_T2 and comp_qual$IgG2_Exces
## X-squared = 2.865, df = 1, p-value = 0.09052
fisher.test(comp_qual$Profil_T2,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$Profil_T2 and comp_qual$IgG2_Exces
## p-value = 0.1364
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.02964217 1.37410003
## sample estimates:
## odds ratio 
##  0.2851939

Réponse biotherapie

xtabs(~Rep_biotherapie_GETE+IgG2_Exces, data=comp_qual)
##                     IgG2_Exces
## Rep_biotherapie_GETE Non Oui
##                      155  26
##          Aggravation   8   1
##          Bonne        25   0
##          Excellent     3   2
##          Faible       25   3
##          modérée       8   2
chisq.test(comp_qual$Rep_biotherapie_GETE,comp_qual$IgG2_Exces)$expected
## Warning in chisq.test(comp_qual$Rep_biotherapie_GETE, comp_qual$IgG2_Exces):
## Chi-squared approximation may be incorrect
##                               comp_qual$IgG2_Exces
## comp_qual$Rep_biotherapie_GETE        Non        Oui
##                                157.147287 23.8527132
##                    Aggravation   7.813953  1.1860465
##                    Bonne        21.705426  3.2945736
##                    Excellent     4.341085  0.6589147
##                    Faible       24.310078  3.6899225
##                    modérée       8.682171  1.3178295
chisq.test(comp_qual$Rep_biotherapie_GETE,comp_qual$IgG2_Exces, correct=FALSE)
## Warning in chisq.test(comp_qual$Rep_biotherapie_GETE, comp_qual$IgG2_Exces, :
## Chi-squared approximation may be incorrect
## 
##  Pearson's Chi-squared test
## 
## data:  comp_qual$Rep_biotherapie_GETE and comp_qual$IgG2_Exces
## X-squared = 7.75, df = 5, p-value = 0.1706
fisher.test(comp_qual$Rep_biotherapie_GETE,comp_qual$IgG2_Exces)
## 
##  Fisher's Exact Test for Count Data
## 
## data:  comp_qual$Rep_biotherapie_GETE and comp_qual$IgG2_Exces
## p-value = 0.09327
## alternative hypothesis: two.sided