该研究检索了Cochrane对照试验中心注册,CINAHL
,Embase
,LILACS数据库
,MEDLINE
,MEDLINEIn-Process
,PsycINFO
,监管机构网站
,以及从一开始就发布和未发表的双盲随机对照试验的国际注册2016年1月8日。
研究包括21种抗抑郁药的安慰剂对照和头对头试验,用于成人(≥18岁和两性)的急性治疗,根据标准操作标准诊断为严重抑郁症。
主要结果是疗效(反应率)和可接受性(由于任何原因导致的治疗中断)。我们使用具有随机效应的成对和网络荟萃分析估计概要比值比(OR)。
数据来源于该文章提供的数据,https://data.mendeley.com/datasets/83rthbp8ys/2
。
1 程序包安装及数据加载
首先是下载和加载netmeta程序包。
##install.packages("netmeta")
library(netmeta)
eff <- read.csv("full-efficacy.csv",
header = TRUE,
stringsAsFactors = FALSE)
acc <- read.csv("full-acceptability.csv",
header = TRUE,
stringsAsFactors = FALSE)
2 模型构建(netmeta)
模型构建首先使用pairwise
函数对数据处理其次鞭尸使用核心函数netmeta
函数进行模型分析,可以分别选择fixed和random两类模型。
deff <- pairwise(treat = t,
n,
event = r,
data = eff,
studlab = id,
sm = "OR")
dacc <- pairwise(treat = t,
n,
event = r,
data = acc,
studlab = id,
sm = "OR")
e.netmeta <-netmeta(deff,
comb.fixed = FALSE,
comb.random = TRUE)
e.netmeta
a.netmeta <-netmeta(dacc,
comb.fixed = FALSE,
comb.random = TRUE)
a.netmeta
模型结果如下所示(仅展示e.netmeta):
## Number of studies: k = 433
## Number of pairwise comparisons: m = 555
## Number of observations: o = 103155
## Number of treatments: n = 22
## Number of designs: d = 117
##
## Random effects model
##
## Treatment estimate (sm = 'OR', comparison: other treatments vs 'agomelatine'):
## OR 95%-CI z p-value
## agomelatine . . . .
## amitriptyline 1.2819 [1.0792; 1.5226] 2.83 0.0047
## bupropion 0.9461 [0.7759; 1.1536] -0.55 0.5841
## citalopram 0.9289 [0.7753; 1.1130] -0.80 0.4239
## clomipramine 0.9061 [0.7092; 1.1577] -0.79 0.4304
## desvenlafaxine 0.9078 [0.7258; 1.1355] -0.85 0.3969
## duloxetine 1.1242 [0.9537; 1.3252] 1.40 0.1629
## escitalopram 1.0180 [0.8664; 1.1963] 0.22 0.8281
## fluoxetine 0.9246 [0.8021; 1.0660] -1.08 0.2803
## fluvoxamine 1.0197 [0.8218; 1.2653] 0.18 0.8593
## levomilnacipran 0.9716 [0.7343; 1.2857] -0.20 0.8404
## milnacipran 1.0531 [0.8032; 1.3808] 0.37 0.7082
## mirtazapine 1.1393 [0.9415; 1.3787] 1.34 0.1802
## nefazodone 1.0155 [0.7763; 1.3284] 0.11 0.9109
## paroxetine 1.0562 [0.9170; 1.2165] 0.76 0.4482
## placebo 0.6118 [0.5373; 0.6967] -7.41 < 0.0001
## reboxetine 0.8301 [0.6729; 1.0240] -1.74 0.0821
## sertraline 1.0080 [0.8516; 1.1932] 0.09 0.9259
## trazodone 0.9050 [0.7191; 1.1389] -0.85 0.3947
## venlafaxine 1.0800 [0.9256; 1.2600] 0.98 0.3282
## vilazodone 0.9636 [0.7475; 1.2421] -0.29 0.7746
## vortioxetine 1.0130 [0.8384; 1.2239] 0.13 0.8938
##
## Quantifying heterogeneity / inconsistency:
## tau^2 = 0.0426; tau = 0.2063; I^2 = 34.1% [26.1%; 41.2%]
##
## Tests of heterogeneity (within designs) and inconsistency (between designs):
## Q d.f. p-value
## Total 717.35 473 < 0.0001
## Within designs 499.41 347 < 0.0001
## Between designs 217.95 126 < 0.0001
3 网络证据图(network plot)
网状证据图是用来表示各种干预措施之间关系的最直观的可视化,是网状meta分析的标配,核心函数为netgraph
。
-
网状meta分析的核心结果之一
-
节点:干预措施的样本量
-
线条:干预措施头对头试验,粗细代表试验数
##有效性 efficacy
netgraph(e.netmeta,
start ="circle",
cex = 1,
col = "black",
plastic = FALSE,
points = TRUE,
col.points = "steelblue",
cex.points =15*sqrt(n.trts/max(n.trts)),
thickness = "number.of.studies",
lwd.max = 12,
lwd.min = 1,
multiarm = F)
netgraph(a.netmeta,
start ="circle",
cex = 0.7,
col = "black",
plastic = F,
points = TRUE,
col.points = "darkred",
cex.points =15*sqrt(n.trts/max(n.trts)),
thickness = "number.of.studies",
lwd.max = 12,
lwd.min = 1,
multiarm = FALSE)
4 森林图(forest plot)
森林图是meta分析的标配,在网状meta中也是一样的,如图所示,森林图给出了各干预措施比较的效应量及95%可信区间,当然,也给出了预测区间。
主要使用forest
函数实现。
forest(e.netmeta,
ref = "placebo",
pooled = "random",
digits = 2,
col.square = "steelblue",
smlab = "Random effects model",
label.left = "Favors citalopram",
label.right = "Favors the other",
leftlabs = "Efficacy",
drop = TRUE,
sortvar = -TE)
forest(a.netmeta,
ref = "placebo",
pooled = "random",
digits = 2,
col.square = "darkred",
xlim = c(2, 0.5),
smlab = "Random effects model",
label.left = "Favors the other",
label.right = "Favors citalopram",
leftlabs = "Acceptability",
drop = T,
sortvar = TE)
6 两两比较赛联表(league)
赛联表为网状meta两两比较结果,使用netleague
函数实现。
league <- netleague(e.netmeta,
a.netmeta,
backtransf = TRUE,
direct = FALSE,
fixed = FALSE,
digits =2,
bracket = "(",
separator = " to ")
league
6 节点分割法(netsplit)
不一致性是指网状meta分析中直接证据和间接证据存在差异,这会影响网状meta分析的真实性,因此需要在进行网状meta分析时进行检测,并分析不一致性的产生原因。对于三个治疗措施形成的闭合环,就可以直接比较直接证据和间接证据之间的不一致性。对于四个研究形成的闭合环,可以分割成两个闭合的三角形环,进而分析直接证据和间接证据之间的不一致性。
常用的不一致检验方法为节点分割法,使用netsplit函数实现:
netsplit(e.netmeta)
## Separate indirect from direct evidence (SIDE) using back-calculation method
##
## Random effects model:
##
## comparison k prop nma direct indir. RoR z p-value
## amitriptyline:agomelatine 0 0 1.2819 . 1.2819 . . .
## bupropion:agomelatine 0 0 0.9461 . 0.9461 . . .
## citalopram:agomelatine 0 0 0.9289 . 0.9289 . . .
## clomipramine:agomelatine 0 0 0.9061 . 0.9061 . . .
## desvenlafaxine:agomelatine 0 0 0.9078 . 0.9078 . . .
## duloxetine:agomelatine 1 0.08 1.1242 1.3162 1.1087 1.1873 0.56 0.5763
## escitalopram:agomelatine 2 0.10 1.0180 0.8117 1.0431 0.7782 -0.90 0.3671
## fluoxetine:agomelatine 4 0.26 0.9246 0.9841 0.9050 1.0874 0.50 0.6144
## fluvoxamine:agomelatine 0 0 1.0197 . 1.0197 . . .
## levomilnacipran:agomelatine 0 0 0.9716 . 0.9716 . . .
## milnacipran:agomelatine 0 0 1.0531 . 1.0531 . . .
## mirtazapine:agomelatine 0 0 1.1393 . 1.1393 . . .
## nefazodone:agomelatine 0 0 1.0155 . 1.0155 . . .
## paroxetine:agomelatine 5 0.28 1.0562 1.1715 1.0151 1.1540 0.89 0.3740
## placebo:agomelatine 13 0.57 0.6118 0.6423 0.5742 1.1186 0.84 0.4022
## reboxetine:agomelatine 0 0 0.8301 . 0.8301 . . .
## sertraline:agomelatine 0 0 1.0080 . 1.0080 . . .
## trazodone:agomelatine 0 0 0.9050 . 0.9050 . . .
## venlafaxine:agomelatine 2 0.10 1.0800 0.7374 1.1293 0.6529 -1.66 0.0970
## vilazodone:agomelatine 0 0 0.9636 . 0.9636 . . .
## vortioxetine:agomelatine 0 0 1.0130 . 1.0130 . . .
## amitriptyline:bupropion 0 0 1.3549 . 1.3549 . . .
## amitriptyline:citalopram 0 0 1.3800 . 1.3800 . . .
## amitriptyline:clomipramine 0 0 1.4147 . 1.4147 . . .
## amitriptyline:desvenlafaxine 0 0 1.4120 . 1.4120 . . .
## amitriptyline:duloxetine 0 0 1.1402 . 1.1402 . . .
## amitriptyline:escitalopram 0 0 1.2592 . 1.2592 . . .
## amitriptyline:fluoxetine 12 0.17 1.3863 1.0645 1.4642 0.7271 -1.78 0.0751
## amitriptyline:fluvoxamine 3 0.15 1.2571 1.0897 1.2886 0.8457 -0.58 0.5605
## amitriptyline:levomilnacipran 0 0 1.3193 . 1.3193 . . .
## amitriptyline:milnacipran 2 0.17 1.2172 1.1673 1.2274 0.9510 -0.14 0.8865
## amitriptyline:mirtazapine 4 0.15 1.1251 1.0455 1.1397 0.9173 -0.34 0.7325
## amitriptyline:nefazodone 0 0 1.2624 . 1.2624 . . .
## amitriptyline:paroxetine 13 0.27 1.2137 1.0677 1.2734 0.8385 -1.19 0.2326
## amitriptyline:placebo 22 0.41 2.0951 2.6250 1.7952 1.4622 3.04 0.0024
## amitriptyline:reboxetine 0 0 1.5442 . 1.5442 . . .
## amitriptyline:sertraline 7 0.31 1.2716 1.2908 1.2632 1.0218 0.13 0.8948
## amitriptyline:trazodone 3 0.19 1.4165 1.2748 1.4511 0.8785 -0.46 0.6458
## amitriptyline:venlafaxine 2 0.06 1.1869 1.1829 1.1872 0.9963 -0.01 0.9905
## amitriptyline:vilazodone 0 0 1.3303 . 1.3303 . . .
## amitriptyline:vortioxetine 0 0 1.2655 . 1.2655 . . .
## bupropion:citalopram 0 0 1.0186 . 1.0186 . . .
## bupropion:clomipramine 0 0 1.0442 . 1.0442 . . .
## bupropion:desvenlafaxine 0 0 1.0422 . 1.0422 . . .
## bupropion:duloxetine 0 0 0.8416 . 0.8416 . . .
## bupropion:escitalopram 2 0.17 0.9294 0.9319 0.9289 1.0033 0.01 0.9894
## bupropion:fluoxetine 3 0.19 1.0232 0.8415 1.0726 0.7846 -1.13 0.2601
## bupropion:fluvoxamine 0 0 0.9278 . 0.9278 . . .
## bupropion:levomilnacipran 0 0 0.9738 . 0.9738 . . .
## bupropion:milnacipran 0 0 0.8984 . 0.8984 . . .
## bupropion:mirtazapine 0 0 0.8304 . 0.8304 . . .
## bupropion:nefazodone 0 0 0.9317 . 0.9317 . . .
## bupropion:paroxetine 1 0.05 0.8958 1.0008 0.8909 1.1234 0.29 0.7743
## bupropion:placebo 17 0.75 1.5464 1.4492 1.8839 0.7693 -1.45 0.1472
## bupropion:reboxetine 0 0 1.1398 . 1.1398 . . .
## bupropion:sertraline 1 < 0.01 0.9386 0.9375 0.9386 0.9988 -0.00 0.9991
## bupropion:trazodone 1 0.08 1.0455 2.0952 0.9811 2.1357 1.72 0.0859
## bupropion:venlafaxine 1 0.09 0.8761 1.1722 0.8510 1.3774 1.03 0.3052
## bupropion:vilazodone 0 0 0.9819 . 0.9819 . . .
## bupropion:vortioxetine 0 0 0.9340 . 0.9340 . . .
## citalopram:clomipramine 1 0.08 1.0251 0.5674 1.0811 0.5248 -1.43 0.1524
## citalopram:desvenlafaxine 0 0 1.0232 . 1.0232 . . .
## citalopram:duloxetine 0 0 0.8263 . 0.8263 . . .
## citalopram:escitalopram 13 0.43 0.9124 0.7465 1.0587 0.7052 -2.28 0.0226
## citalopram:fluoxetine 2 0.11 1.0046 0.9632 1.0101 0.9536 -0.20 0.8390
## citalopram:fluvoxamine 1 0.09 0.9109 1.1071 0.8937 1.2388 0.56 0.5728
## citalopram:levomilnacipran 0 0 0.9560 . 0.9560 . . .
## citalopram:milnacipran 0 0 0.8820 . 0.8820 . . .
## citalopram:mirtazapine 1 0.06 0.8153 1.3238 0.7926 1.6702 1.21 0.2263
## citalopram:nefazodone 0 0 0.9148 . 0.9148 . . .
## citalopram:paroxetine 1 0.06 0.8795 1.3555 0.8575 1.5808 1.39 0.1631
## citalopram:placebo 11 0.47 1.5182 1.5025 1.5324 0.9805 -0.15 0.8833
## citalopram:reboxetine 1 0.12 1.1190 1.7296 1.0517 1.6446 1.57 0.1167
## citalopram:sertraline 3 0.11 0.9215 0.8849 0.9259 0.9557 -0.17 0.8682
## citalopram:trazodone 0 0 1.0264 . 1.0264 . . .
## citalopram:venlafaxine 2 0.05 0.8601 1.7421 0.8265 2.1078 2.09 0.0365
## citalopram:vilazodone 2 0.38 0.9640 0.9358 0.9813 0.9536 -0.19 0.8516
## citalopram:vortioxetine 0 0 0.9170 . 0.9170 . . .
## clomipramine:desvenlafaxine 0 0 0.9981 . 0.9981 . . .
## clomipramine:duloxetine 0 0 0.8060 . 0.8060 . . .
## clomipramine:escitalopram 0 0 0.8901 . 0.8901 . . .
## clomipramine:fluoxetine 4 0.19 0.9800 0.6108 1.0955 0.5576 -2.07 0.0389
## clomipramine:fluvoxamine 2 0.06 0.8886 1.8541 0.8450 2.1942 1.41 0.1572
## clomipramine:levomilnacipran 0 0 0.9326 . 0.9326 . . .
## clomipramine:milnacipran 1 0.12 0.8604 0.8696 0.8592 1.0121 0.03 0.9798
## clomipramine:mirtazapine 0 0 0.7953 . 0.7953 . . .
## clomipramine:nefazodone 0 0 0.8923 . 0.8923 . . .
## clomipramine:paroxetine 6 0.44 0.8579 0.9052 0.8227 1.1003 0.44 0.6617
## clomipramine:placebo 0 0 1.4810 . 1.4810 . . .
## clomipramine:reboxetine 0 0 1.0916 . 1.0916 . . .
## clomipramine:sertraline 2 0.16 0.8989 0.9547 0.8883 1.0746 0.23 0.8204
## clomipramine:trazodone 1 0.10 1.0013 1.6162 0.9476 1.7056 1.14 0.2524
## clomipramine:venlafaxine 2 0.13 0.8390 0.6046 0.8797 0.6872 -1.08 0.2811
## clomipramine:vilazodone 0 0 0.9404 . 0.9404 . . .
## clomipramine:vortioxetine 0 0 0.8945 . 0.8945 . . .
## desvenlafaxine:duloxetine 1 0.14 0.8075 0.8285 0.8041 1.0304 0.10 0.9223
## desvenlafaxine:escitalopram 0 0 0.8917 . 0.8917 . . .
## desvenlafaxine:fluoxetine 0 0 0.9818 . 0.9818 . . .
## desvenlafaxine:fluvoxamine 0 0 0.8903 . 0.8903 . . .
## desvenlafaxine:levomilnacipran 0 0 0.9343 . 0.9343 . . .
## desvenlafaxine:milnacipran 0 0 0.8620 . 0.8620 . . .
## desvenlafaxine:mirtazapine 0 0 0.7968 . 0.7968 . . .
## desvenlafaxine:nefazodone 0 0 0.8940 . 0.8940 . . .
## desvenlafaxine:paroxetine 0 0 0.8595 . 0.8595 . . .
## desvenlafaxine:placebo 9 0.96 1.4837 1.4711 1.8067 0.8142 -0.44 0.6590
## desvenlafaxine:reboxetine 0 0 1.0936 . 1.0936 . . .
## desvenlafaxine:sertraline 0 0 0.9006 . 0.9006 . . .
## desvenlafaxine:trazodone 0 0 1.0031 . 1.0031 . . .
## desvenlafaxine:venlafaxine 0 0 0.8406 . 0.8406 . . .
## desvenlafaxine:vilazodone 0 0 0.9421 . 0.9421 . . .
## desvenlafaxine:vortioxetine 0 0 0.8962 . 0.8962 . . .
## duloxetine:escitalopram 3 0.19 1.1043 0.8436 1.1744 0.7183 -1.73 0.0834
## duloxetine:fluoxetine 2 0.04 1.2158 1.2298 1.2152 1.0120 0.04 0.9713
## duloxetine:fluvoxamine 0 0 1.1025 . 1.1025 . . .
## duloxetine:levomilnacipran 0 0 1.1570 . 1.1570 . . .
## duloxetine:milnacipran 0 0 1.0675 . 1.0675 . . .
## duloxetine:mirtazapine 0 0 0.9868 . 0.9868 . . .
## duloxetine:nefazodone 0 0 1.1071 . 1.1071 . . .
## duloxetine:paroxetine 7 0.28 1.0644 0.9756 1.1018 0.8855 -0.85 0.3975
## duloxetine:placebo 21 0.65 1.8374 1.9167 1.6999 1.1275 1.01 0.3148
## duloxetine:reboxetine 0 0 1.3543 . 1.3543 . . .
## duloxetine:sertraline 0 0 1.1153 . 1.1153 . . .
## duloxetine:trazodone 0 0 1.2423 . 1.2423 . . .
## duloxetine:venlafaxine 2 0.12 1.0410 0.8989 1.0623 0.8462 -0.76 0.4462
## duloxetine:vilazodone 0 0 1.1667 . 1.1667 . . .
## duloxetine:vortioxetine 6 0.46 1.1098 1.3793 0.9188 1.5011 2.46 0.0141
## escitalopram:fluoxetine 3 0.11 1.1010 1.1260 1.0979 1.0256 0.12 0.9028
## escitalopram:fluvoxamine 0 0 0.9983 . 0.9983 . . .
## escitalopram:levomilnacipran 0 0 1.0478 . 1.0478 . . .
## escitalopram:milnacipran 0 0 0.9667 . 0.9667 . . .
## escitalopram:mirtazapine 0 0 0.8936 . 0.8936 . . .
## escitalopram:nefazodone 0 0 1.0025 . 1.0025 . . .
## escitalopram:paroxetine 3 0.14 0.9639 0.9924 0.9594 1.0344 0.18 0.8556
## escitalopram:placebo 20 0.57 1.6639 1.4412 2.0122 0.7162 -3.00 0.0027
## escitalopram:reboxetine 0 0 1.2264 . 1.2264 . . .
## escitalopram:sertraline 3 0.11 1.0099 0.8644 1.0296 0.8395 -0.72 0.4724
## escitalopram:trazodone 0 0 1.1249 . 1.1249 . . .
## escitalopram:venlafaxine 2 0.08 0.9426 1.2105 0.9219 1.3130 1.06 0.2899
## escitalopram:vilazodone 0 0 1.0565 . 1.0565 . . .
## escitalopram:vortioxetine 0 0 1.0050 . 1.0050 . . .
## fluoxetine:fluvoxamine 2 0.11 0.9068 0.9700 0.8992 1.0788 0.25 0.8014
## fluoxetine:levomilnacipran 0 0 0.9516 . 0.9516 . . .
## fluoxetine:milnacipran 2 0.27 0.8780 1.1789 0.7874 1.4972 1.43 0.1517
## fluoxetine:mirtazapine 6 0.22 0.8116 0.7559 0.8281 0.9128 -0.48 0.6312
## fluoxetine:nefazodone 3 0.14 0.9106 0.9633 0.9025 1.0673 0.18 0.8586
## fluoxetine:paroxetine 12 0.23 0.8755 0.9421 0.8566 1.0998 0.79 0.4297
## fluoxetine:placebo 39 0.47 1.5113 1.4097 1.6086 0.8764 -1.56 0.1180
## fluoxetine:reboxetine 4 0.25 1.1139 1.2160 1.0816 1.1243 0.56 0.5785
## fluoxetine:sertraline 6 0.19 0.9173 0.7004 0.9786 0.7157 -2.03 0.0426
## fluoxetine:trazodone 4 0.12 1.0217 0.8357 1.0506 0.7955 -0.73 0.4683
## fluoxetine:venlafaxine 14 0.33 0.8562 0.8008 0.8849 0.9049 -0.82 0.4097
## fluoxetine:vilazodone 0 0 0.9596 . 0.9596 . . .
## fluoxetine:vortioxetine 0 0 0.9128 . 0.9128 . . .
## fluvoxamine:levomilnacipran 0 0 1.0495 . 1.0495 . . .
## fluvoxamine:milnacipran 2 0.21 0.9683 0.5832 1.1054 0.5276 -1.82 0.0690
## fluvoxamine:mirtazapine 2 0.19 0.8950 0.8764 0.8993 0.9745 -0.09 0.9264
## fluvoxamine:nefazodone 0 0 1.0042 . 1.0042 . . .
## fluvoxamine:paroxetine 2 0.08 0.9655 0.8810 0.9729 0.9055 -0.28 0.7787
## fluvoxamine:placebo 11 0.37 1.6667 2.0337 1.4836 1.3708 1.69 0.0915
## fluvoxamine:reboxetine 0 0 1.2284 . 1.2284 . . .
## fluvoxamine:sertraline 2 0.09 1.0116 1.5585 0.9685 1.6091 1.34 0.1802
## fluvoxamine:trazodone 0 0 1.1268 . 1.1268 . . .
## fluvoxamine:venlafaxine 1 0.05 0.9442 0.4229 0.9810 0.4311 -1.76 0.0792
## fluvoxamine:vilazodone 0 0 1.0583 . 1.0583 . . .
## fluvoxamine:vortioxetine 0 0 1.0067 . 1.0067 . . .
## levomilnacipran:milnacipran 0 0 0.9226 . 0.9226 . . .
## levomilnacipran:mirtazapine 0 0 0.8528 . 0.8528 . . .
## levomilnacipran:nefazodone 0 0 0.9568 . 0.9568 . . .
## levomilnacipran:paroxetine 0 0 0.9199 . 0.9199 . . .
## levomilnacipran:placebo 5 1.00 1.5881 1.5881 . . . .
## levomilnacipran:reboxetine 0 0 1.1705 . 1.1705 . . .
## levomilnacipran:sertraline 0 0 0.9639 . 0.9639 . . .
## levomilnacipran:trazodone 0 0 1.0737 . 1.0737 . . .
## levomilnacipran:venlafaxine 0 0 0.8997 . 0.8997 . . .
## levomilnacipran:vilazodone 0 0 1.0084 . 1.0084 . . .
## levomilnacipran:vortioxetine 0 0 0.9592 . 0.9592 . . .
## milnacipran:mirtazapine 0 0 0.9243 . 0.9243 . . .
## milnacipran:nefazodone 0 0 1.0371 . 1.0371 . . .
## milnacipran:paroxetine 2 0.40 0.9971 0.9517 1.0283 0.9255 -0.31 0.7587
## milnacipran:placebo 0 0 1.7212 . 1.7212 . . .
## milnacipran:reboxetine 0 0 1.2686 . 1.2686 . . .
## milnacipran:sertraline 1 0.02 1.0447 2.0870 1.0299 2.0264 0.75 0.4556
## milnacipran:trazodone 0 0 1.1637 . 1.1637 . . .
## milnacipran:venlafaxine 0 0 0.9751 . 0.9751 . . .
## milnacipran:vilazodone 0 0 1.0929 . 1.0929 . . .
## milnacipran:vortioxetine 0 0 1.0396 . 1.0396 . . .
## mirtazapine:nefazodone 0 0 1.1220 . 1.1220 . . .
## mirtazapine:paroxetine 5 0.22 1.0787 0.9653 1.1127 0.8676 -0.75 0.4553
## mirtazapine:placebo 12 0.34 1.8621 1.7805 1.9058 0.9342 -0.43 0.6659
## mirtazapine:reboxetine 0 0 1.3725 . 1.3725 . . .
## mirtazapine:sertraline 1 0.08 1.1302 0.9741 1.1458 0.8502 -0.50 0.6138
## mirtazapine:trazodone 2 0.18 1.2589 1.5017 1.2127 1.2384 0.70 0.4859
## mirtazapine:venlafaxine 2 0.14 1.0549 1.3661 1.0124 1.3493 1.22 0.2208
## mirtazapine:vilazodone 0 0 1.1824 . 1.1824 . . .
## mirtazapine:vortioxetine 0 0 1.1247 . 1.1247 . . .
## nefazodone:paroxetine 2 0.16 0.9614 0.7517 1.0065 0.7468 -0.85 0.3962
## nefazodone:placebo 8 0.63 1.6597 1.7354 1.5368 1.1293 0.48 0.6285
## nefazodone:reboxetine 0 0 1.2233 . 1.2233 . . .
## nefazodone:sertraline 1 0.12 1.0074 1.1667 0.9874 1.1816 0.41 0.6789
## nefazodone:trazodone 0 0 1.1221 . 1.1221 . . .
## nefazodone:venlafaxine 0 0 0.9403 . 0.9403 . . .
## nefazodone:vilazodone 0 0 1.0538 . 1.0538 . . .
## nefazodone:vortioxetine 0 0 1.0025 . 1.0025 . . .
## paroxetine:placebo 46 0.53 1.7263 1.6524 1.8138 0.9110 -1.14 0.2554
## paroxetine:reboxetine 3 0.31 1.2724 1.2385 1.2881 0.9615 -0.20 0.8411
## paroxetine:sertraline 2 0.08 1.0478 0.8952 1.0620 0.8429 -0.70 0.4867
## paroxetine:trazodone 2 0.12 1.1671 1.7157 1.1093 1.5467 1.36 0.1739
## paroxetine:venlafaxine 3 0.09 0.9780 0.6303 1.0189 0.6186 -2.25 0.0243
## paroxetine:vilazodone 0 0 1.0961 . 1.0961 . . .
## paroxetine:vortioxetine 0 0 1.0427 . 1.0427 . . .
## placebo:reboxetine 10 0.62 0.7371 0.7123 0.7788 0.9145 -0.50 0.6157
## placebo:sertraline 18 0.47 0.6070 0.6270 0.5898 1.0630 0.52 0.6009
## placebo:trazodone 8 0.42 0.6761 0.5699 0.7653 0.7446 -1.47 0.1410
## placebo:venlafaxine 22 0.46 0.5665 0.5614 0.5709 0.9833 -0.17 0.8668
## placebo:vilazodone 6 0.90 0.6350 0.6472 0.5318 1.2171 0.52 0.6024
## placebo:vortioxetine 14 0.83 0.6040 0.5766 0.7538 0.7649 -1.42 0.1565
## reboxetine:sertraline 0 0 0.8235 . 0.8235 . . .
## reboxetine:trazodone 0 0 0.9173 . 0.9173 . . .
## reboxetine:venlafaxine 1 0.07 0.7686 1.0446 0.7519 1.3894 0.85 0.3956
## reboxetine:vilazodone 0 0 0.8615 . 0.8615 . . .
## reboxetine:vortioxetine 0 0 0.8195 . 0.8195 . . .
## sertraline:trazodone 1 0.06 1.1139 0.5595 1.1659 0.4799 -1.60 0.1095
## sertraline:venlafaxine 3 0.10 0.9334 0.8472 0.9436 0.8979 -0.45 0.6520
## sertraline:vilazodone 0 0 1.0461 . 1.0461 . . .
## sertraline:vortioxetine 0 0 0.9951 . 0.9951 . . .
## trazodone:venlafaxine 2 0.13 0.8379 0.6311 0.8743 0.7218 -1.03 0.3052
## trazodone:vilazodone 0 0 0.9392 . 0.9392 . . .
## trazodone:vortioxetine 0 0 0.8934 . 0.8934 . . .
## venlafaxine:vilazodone 0 0 1.1208 . 1.1208 . . .
## venlafaxine:vortioxetine 2 0.15 1.0661 0.8452 1.1114 0.7605 -1.17 0.2412
## vilazodone:vortioxetine 0 0 0.9512 . 0.9512 . . .
##
## Legend:
## comparison - Treatment comparison
## k - Number of studies providing direct evidence
## prop - Direct evidence proportion
## nma - Estimated treatment effect (OR) in network meta-analysis
## direct - Estimated treatment effect (OR) derived from direct evidence
## indir. - Estimated treatment effect (OR) derived from indirect evidence
## RoR - Ratio of Ratios (direct versus indirect)
## z - z-value of test for disagreement (direct versus indirect)
## p-value - p-value of test for disagreement (direct versus indirect)
forest(netsplit(e.netmeta))
结果同样可以使用森林图展示(展示部分):
7 热图(netheat plot)
-
一种判断不一致性的方法,图中的灰色方块在row方向中越大表示研究越重要,颜色越红代表研究一致性越大。
netheat(e.netmeta,
nchar.trts = 4,
random = TRUE)
8 排序图(SURCA plot)
对干预措施的优劣进行排序是网状meta分析的一大特色,也是一个主要的优势。目前排序的方法比较多,最常用的有SUCRA法和P得分法,其中,SUCRA法是最为常用的方法,该值越大为最佳干预措施的可能性越大。
netrank(e.netmeta, small.values = "bad") #可做条形图
## P-score
## amitriptyline 0.9840
## mirtazapine 0.8460
## duloxetine 0.8367
## venlafaxine 0.7506
## paroxetine 0.6937
## milnacipran 0.6427
## escitalopram 0.5799
## fluvoxamine 0.5766
## nefazodone 0.5608
## vortioxetine 0.5606
## sertraline 0.5493
## agomelatine 0.5248
## levomilnacipran 0.4607
## vilazodone 0.4374
## bupropion 0.3761
## citalopram 0.3207
## clomipramine 0.2962
## fluoxetine 0.2939
## desvenlafaxine 0.2908
## trazodone 0.2865
## reboxetine 0.1318
## placebo 0.0000
9 漏斗图(funnel plot)
发表偏倚检测实际上是和传统meta分析一致的,主要靠漏斗图来实现。解读方法看是否对称,是一个主观判断。
colors <- c(
agom = "thistle", amit = "lightgreen",
bupr = "coral3", cita = "cadetblue4",
clom = "orange3",esci = "pink",
fluv = "dodgerblue4", miln = "goldenrod4",
mirt = "yellow3", nefa = "darkgrey",
paro = "gray9", rebo = "lightblue3",
sert = "lightslateblue", traz = "red4",
venl= "mediumvioletred")
trts_ef <- substr(e.netmeta$trts, 0, 4)
trts_ac <- substr(a.netmeta$trts, 0, 4)
comparison_ef<- trts_ef[trts_ef != "fluo"]
comparison_ac <- trts_ac[trts_ac != "fluo"]
ord_ef<- c(comparison_ef, "fluo")
ord_ac <- c(comparison_ac, "fluo")
然后使用funnel
函数进行漏斗图进行绘制:
funnel(netmeta_ef,
order = ord_ef,
pch = rep(19),
col = colors,
legend = FALSE,
linreg = TRUE,
text.linreg = "(Egger's test)",
pos.tests = "topright")
legend("topleft",
legend = comparison_ef,
pch = rep(19),
col = colors,
cex = 0.75)
funnel(
netmeta_ac,
order = ord_ac,
pch = rep(19),
col = colors,
legend = FALSE,
linreg = TRUE,
text.linreg = "(Egger's test)",
pos.tests = "topright")
legend("topleft",
legend = comparison_ac,
pch = rep(19),
col = colors,
cex = 0.75)