Valores de Referência de Espirometria em Crianças Brasileiras

Author

Marcus Herbert Jones

Published

May 3, 2026

Setup

Code 
library(tidyverse)
library(readxl)
library(ggpubr)
library(expss)
library(ggsci)
library(quantreg)
library(patchwork)
library(ggcorrplot)
library(gamlss)
library(gamlss.add)
library(gamlss.dist)
library(writexl)
library(haven)
Code 
palette_nejm <- c(
  "#BC3C29FF", "#0072B5FF", "#E18727FF", "#20854EFF",
  "#7876B1FF", "#6F99ADFF", "#FFDC91FF", "#EE4C97FF"
)
palette_nejmb <- c(
  "#0072B5FF", "#EE4C97FF", "#20854EFF", "#7876B1FF",
  "#6F99ADFF", "#FFDC91FF", "#E18727FF", "#BC3C29FF"
)

Carregamento e Preparação dos Dados

Code 
x2005 <- read_excel("2005f.xlsx")

# Calcular FEF25-75/CVF
x2005$feffvc <- x2005$fef2575 / x2005$fvc

# Converter variáveis categóricas e padronizar rótulos de etnia
x2005 <- x2005 |>
  mutate(
    sexf     = as.factor(sexf),
    ethnic01 = as.factor(ethnic01),
    ethnic   = as.factor(ethnic),
    protocol = as.factor(protocol),
    ethnicf  = case_when(
      ethnicf == "Branco"        ~ "Branca",
      ethnicf == "Negro e Pardo" ~ "Negra e Parda",
      TRUE                       ~ ethnicf
    ) |> as.factor(),
    ethnicnpb = case_when(
      ethnic == 1 ~ "Branca",
      ethnic == 2 ~ "Negra",
      ethnic == 5 ~ "Parda"
    ) |> as.factor()
  )

Estatísticas Descritivas

Code 
x2005 |> count(sexf)
# A tibble: 2 × 2
  sexf          n
  <fct>     <int>
1 Feminino   1059
2 Masculino   931
Code 
x2005 |> count(ethnic)
# A tibble: 4 × 2
  ethnic     n
  <fct>  <int>
1 1       1353
2 2        184
3 5        386
4 <NA>      67
Code 
x2005 |> count(protocol, sex, ethnic)
# A tibble: 14 × 4
   protocol   sex ethnic     n
   <fct>    <dbl> <fct>  <int>
 1 1            1 1        318
 2 1            1 2         48
 3 1            1 5         71
 4 1            2 1        366
 5 1            2 2         55
 6 1            2 5         78
 7 2            1 1        313
 8 2            1 2         34
 9 2            1 5        118
10 2            1 <NA>      29
11 2            2 1        356
12 2            2 2         47
13 2            2 5        119
14 2            2 <NA>      38
Code 
x2005 |> count(sexf, ethnic01)
# A tibble: 6 × 3
  sexf      ethnic01     n
  <fct>     <fct>    <int>
1 Feminino  0          722
2 Feminino  1          299
3 Feminino  <NA>        38
4 Masculino 0          631
5 Masculino 1          271
6 Masculino <NA>        29

Tabela 1 — Participantes por Protocolo e Cidade

Code 
table1 <- x2005 |>
  group_by(protocol, cidade) |>
  summarise(count = n(), .groups = "drop") |>
  arrange(protocol, desc(count))

table1 |> htmlTable()
Table 1
protocol cidade count
1 1 Porto Alegre 224
2 1 Recife 176
3 1 Sete Lagoas 164
4 1 São Luiz Gonzaga 76
5 1 Foz do Iguacu 48
6 1 Blumenau 42
7 1 Salvador 42
8 1 Campinas 34
9 1 Ribeirao Preto 32
10 1 Rio de Janeiro 32
11 1 Niteroi 28
12 1 Curitiba 22
13 1 Belo Horizonte 16
14 2 Porto Alegre 555
15 2 São Paulo 266
16 2 Rio Grande 188
17 2 Caxias do Sul 45

Subconjuntos

Code 
# Com dados de etnia
x2005x <- x2005 |> filter(!is.na(ethnic))

# Por sexo (amostra completa)
mm <- x2005 |> filter(sex == 1)
ff <- x2005 |> filter(sex == 2)

# Por sexo, apenas com dados de etnia
mmx <- x2005x |> filter(sex == 1)
ffx <- x2005x |> filter(sex == 2)

Base Objects para os Gráficos

Code 
# Masculino — por etnia
VEF1m <- ggplot(mmx, aes(height, fev1))
CVFm  <- ggplot(mmx, aes(height, fvc))

# Feminino — por etnia
VEF1f <- ggplot(ffx, aes(height, fev1))
CVFf  <- ggplot(ffx, aes(height, fvc))

# Com cor por etnia
VEF1me <- ggplot(mmx, aes(height, fev1, colour = ethnicf))
CVFme  <- ggplot(mmx, aes(height, fvc,  colour = ethnicf))
VEF1fe <- ggplot(ffx, aes(height, fev1, colour = ethnicf))
CVFfe  <- ggplot(ffx, aes(height, fvc,  colour = ethnicf))

# Amostra completa — estratificada por sexo
VEF1s   <- ggplot(x2005x, aes(height, fev1,    colour = sexf))
CVFs    <- ggplot(x2005x, aes(height, fvc,     colour = sexf))
TIFFENAUs <- ggplot(x2005, aes(height, fev1fvc, colour = sexf))
FEF2575s  <- ggplot(x2005, aes(height, fef2575, colour = sexf))
FEFFVCs   <- ggplot(x2005, aes(height, feffvc,  colour = sexf))

# Amostra completa — estratificada por etnia
VEF1e   <- ggplot(x2005x, aes(height, fev1,    colour = ethnicf))
CVFe    <- ggplot(x2005x, aes(height, fvc,     colour = ethnicf))
TIFFENAUe <- ggplot(x2005, aes(height, fev1fvc, colour = ethnicf))
FEF2575e  <- ggplot(x2005, aes(height, fef2575, colour = ethnicf))
FEFFVCe   <- ggplot(x2005, aes(height, feffvc,  colour = ethnicf))

# Sem estratificação
TIFFENAU <- ggplot(x2005, aes(height, fev1fvc))
FEF2575  <- ggplot(x2005, aes(height, fef2575))

Regressões Quantílicas (P50 e P5)

Sexo Masculino

Code 
modfev1mm    <- rq(lnfev1    ~ lnheight + lnage + ethnic01, data = mmx, tau = c(0.5, 0.05))
modfvcmm     <- rq(lnfvc     ~ lnheight + lnage + ethnic01, data = mmx, tau = c(0.75, 0.5, 0.05))
modfev1fvcmm <- rq(fev1fvc   ~ lnheight,                    data = mm,  tau = c(0.5, 0.05))
modfefmm     <- rq(lnfef2575 ~ lnheight + lnage,            data = mm,  tau = c(0.5, 0.05))
modfeffvcmm  <- rq(feffvc    ~ lnheight + ethnic01,          data = mm,  tau = c(0.5, 0.05))

summary(modfev1mm)

Call: rq(formula = lnfev1 ~ lnheight + lnage + ethnic01, tau = c(0.5, 
    0.05), data = mmx)

tau: [1] 0.05

Coefficients:
            coefficients lower bd  upper bd 
(Intercept)  -8.96018    -10.13539  -8.36167
lnheight      1.82020      1.67333   2.08426
lnage         0.20898      0.12530   0.26928
ethnic011    -0.02638     -0.07219  -0.01189

Call: rq(formula = lnfev1 ~ lnheight + lnage + ethnic01, tau = c(0.5, 
    0.05), data = mmx)

tau: [1] 0.5

Coefficients:
            coefficients lower bd  upper bd 
(Intercept) -10.43423    -11.10533  -9.78236
lnheight      2.20761      2.05752   2.37066
lnage         0.11400      0.04048   0.16857
ethnic011    -0.03184     -0.05477  -0.00888
Code 
summary(modfvcmm)

Call: rq(formula = lnfvc ~ lnheight + lnage + ethnic01, tau = c(0.75, 
    0.5, 0.05), data = mmx)

tau: [1] 0.05

Coefficients:
            coefficients lower bd upper bd
(Intercept) -9.32749     -9.65307 -8.08025
lnheight     1.89018      1.57641  1.98713
lnage        0.26892      0.16604  0.41522
ethnic011   -0.04238     -0.12322 -0.01624

Call: rq(formula = lnfvc ~ lnheight + lnage + ethnic01, tau = c(0.75, 
    0.5, 0.05), data = mmx)

tau: [1] 0.5

Coefficients:
            coefficients lower bd  upper bd 
(Intercept) -11.35819    -12.03562 -10.60504
lnheight      2.41982      2.23702   2.57293
lnage         0.11548      0.05253   0.19788
ethnic011    -0.02874     -0.05973  -0.01165

Call: rq(formula = lnfvc ~ lnheight + lnage + ethnic01, tau = c(0.75, 
    0.5, 0.05), data = mmx)

tau: [1] 0.75

Coefficients:
            coefficients lower bd  upper bd 
(Intercept) -11.82284    -12.95717 -10.80870
lnheight      2.55270      2.31368   2.83953
lnage         0.06855     -0.01628   0.14773
ethnic011    -0.02779     -0.04638  -0.01014
Code 
summary(modfev1fvcmm)

Call: rq(formula = fev1fvc ~ lnheight, tau = c(0.5, 0.05), data = mm)

tau: [1] 0.05

Coefficients:
            coefficients lower bd upper bd
(Intercept)  1.94278      1.63377  2.11490
lnheight    -0.23126     -0.26699 -0.16792

Call: rq(formula = fev1fvc ~ lnheight, tau = c(0.5, 0.05), data = mm)

tau: [1] 0.5

Coefficients:
            coefficients lower bd upper bd
(Intercept)  1.76096      1.64941  1.97617
lnheight    -0.17675     -0.22062 -0.15401
Code 
summary(modfefmm)

Call: rq(formula = lnfef2575 ~ lnheight + lnage, tau = c(0.5, 0.05), 
    data = mm)

tau: [1] 0.05

Coefficients:
            coefficients lower bd upper bd
(Intercept) -5.35627     -9.50945 -2.81863
lnheight     1.08515      0.43035  2.12647
lnage        0.21102     -0.21392  0.48229

Call: rq(formula = lnfef2575 ~ lnheight + lnage, tau = c(0.5, 0.05), 
    data = mm)

tau: [1] 0.5

Coefficients:
            coefficients lower bd upper bd
(Intercept) -6.95786     -8.18446 -5.13957
lnheight     1.52730      1.09681  1.78279
lnage        0.13865      0.05500  0.26662
Code 
summary(modfeffvcmm)

Call: rq(formula = feffvc ~ lnheight + ethnic01, tau = c(0.5, 0.05), 
    data = mm)

tau: [1] 0.05

Coefficients:
            coefficients lower bd upper bd
(Intercept)  4.57902      3.26658  5.62751
lnheight    -0.78661     -1.00030 -0.53760
ethnic011    0.03697     -0.06494  0.08569

Call: rq(formula = feffvc ~ lnheight + ethnic01, tau = c(0.5, 0.05), 
    data = mm)

tau: [1] 0.5

Coefficients:
            coefficients lower bd upper bd
(Intercept)  6.18550      5.20489  6.93684
lnheight    -1.04420     -1.19977 -0.84383
ethnic011    0.09330      0.05033  0.13009

Sexo Feminino

Code 
modfev1ff    <- rq(lnfev1    ~ lnheight + lnage + ethnic01, data = ffx, tau = c(0.5, 0.05))
modfvcff     <- rq(lnfvc     ~ lnheight + lnage + ethnic01, data = ffx, tau = c(0.75, 0.5, 0.05))
modfev1fvcff <- rq(fev1fvc   ~ lnheight,                    data = ff,  tau = c(0.5, 0.05))
modfefff     <- rq(lnfef2575 ~ lnheight + lnage,            data = ff,  tau = c(0.5, 0.05))
modfeffvcff  <- rq(feffvc    ~ lnheight + ethnic01,          data = ff,  tau = c(0.5, 0.05))

summary(modfev1ff)

Call: rq(formula = lnfev1 ~ lnheight + lnage + ethnic01, tau = c(0.5, 
    0.05), data = ffx)

tau: [1] 0.05

Coefficients:
            Value    Std. Error t value  Pr(>|t|)
(Intercept) -8.93402  1.03584   -8.62494  0.00000
lnheight     1.76804  0.25067    7.05335  0.00000
lnage        0.29073  0.09855    2.95024  0.00325
ethnic011   -0.05470  0.02847   -1.92099  0.05502

Call: rq(formula = lnfev1 ~ lnheight + lnage + ethnic01, tau = c(0.5, 
    0.05), data = ffx)

tau: [1] 0.5

Coefficients:
            Value     Std. Error t value   Pr(>|t|) 
(Intercept)  -9.96774   0.40176  -24.81003   0.00000
lnheight      2.08895   0.09713   21.50691   0.00000
lnage         0.15039   0.03870    3.88621   0.00011
ethnic011    -0.04858   0.01062   -4.57370   0.00001
Code 
summary(modfvcff)

Call: rq(formula = lnfvc ~ lnheight + lnage + ethnic01, tau = c(0.75, 
    0.5, 0.05), data = ffx)

tau: [1] 0.05

Coefficients:
            Value     Std. Error t value   Pr(>|t|) 
(Intercept) -10.32588   0.82580  -12.50417   0.00000
lnheight      2.09870   0.19893   10.54981   0.00000
lnage         0.23022   0.07921    2.90636   0.00374
ethnic011    -0.07369   0.02529   -2.91342   0.00365

Call: rq(formula = lnfvc ~ lnheight + lnage + ethnic01, tau = c(0.75, 
    0.5, 0.05), data = ffx)

tau: [1] 0.5

Coefficients:
            Value     Std. Error t value   Pr(>|t|) 
(Intercept) -10.74193   0.46365  -23.16812   0.00000
lnheight      2.26198   0.11203   20.19077   0.00000
lnage         0.15684   0.04410    3.55614   0.00039
ethnic011    -0.04101   0.01360   -3.01651   0.00262

Call: rq(formula = lnfvc ~ lnheight + lnage + ethnic01, tau = c(0.75, 
    0.5, 0.05), data = ffx)

tau: [1] 0.75

Coefficients:
            Value     Std. Error t value   Pr(>|t|) 
(Intercept) -10.57651   0.50318  -21.01923   0.00000
lnheight      2.24407   0.12157   18.45839   0.00000
lnage         0.16611   0.04753    3.49477   0.00049
ethnic011    -0.03499   0.01263   -2.77017   0.00570
Code 
summary(modfev1fvcff)

Call: rq(formula = fev1fvc ~ lnheight, tau = c(0.5, 0.05), data = ff)

tau: [1] 0.05

Coefficients:
            Value    Std. Error t value  Pr(>|t|)
(Intercept)  1.31315  0.14111    9.30602  0.00000
lnheight    -0.10080  0.02855   -3.53023  0.00043

Call: rq(formula = fev1fvc ~ lnheight, tau = c(0.5, 0.05), data = ff)

tau: [1] 0.5

Coefficients:
            Value    Std. Error t value  Pr(>|t|)
(Intercept)  1.43768  0.08328   17.26247  0.00000
lnheight    -0.10621  0.01703   -6.23828  0.00000
Code 
summary(modfefff)

Call: rq(formula = lnfef2575 ~ lnheight + lnage, tau = c(0.5, 0.05), 
    data = ff)

tau: [1] 0.05

Coefficients:
            coefficients lower bd  upper bd 
(Intercept)  -7.85519    -10.11875  -4.84476
lnheight      1.67343      0.93487   2.31428
lnage         0.05613     -0.19595   0.40374

Call: rq(formula = lnfef2575 ~ lnheight + lnage, tau = c(0.5, 0.05), 
    data = ff)

tau: [1] 0.5

Coefficients:
            coefficients lower bd upper bd
(Intercept) -7.38547     -8.43694 -6.60596
lnheight     1.59403      1.40070  1.86019
lnage        0.20358      0.09347  0.28041
Code 
summary(modfeffvcff)

Call: rq(formula = feffvc ~ lnheight + ethnic01, tau = c(0.5, 0.05), 
    data = ff)

tau: [1] 0.05

Coefficients:
            coefficients lower bd upper bd
(Intercept)  3.91939      2.05014  5.80472
lnheight    -0.64261     -1.03824 -0.26121
ethnic011    0.02763     -0.03038  0.08347

Call: rq(formula = feffvc ~ lnheight + ethnic01, tau = c(0.5, 0.05), 
    data = ff)

tau: [1] 0.5

Coefficients:
            coefficients lower bd upper bd
(Intercept)  5.04175      4.44641  5.90341
lnheight    -0.78577     -0.96336 -0.64930
ethnic011    0.04647      0.00349  0.08007

Regressão Linear Múltipla

Sexo Masculino

Code 
lm_modfev1mm    <- lm(lnfev1    ~ lnheight + lnage + ethnic01, data = mmx)
lm_modfvcmm     <- lm(lnfvc     ~ lnheight + lnage + ethnic01, data = mmx)
lm_modfev1fvcmm <- lm(fev1fvc   ~ lnheight,                    data = mm)
lm_modfefmm     <- lm(lnfef2575 ~ lnheight + lnage,            data = mm)
lm_modfeffvcmm  <- lm(feffvc    ~ lnheight + ethnic01,          data = mm)

summary(lm_modfev1mm)

Call:
lm(formula = lnfev1 ~ lnheight + lnage + ethnic01, data = mmx)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.54946 -0.08492  0.00029  0.09293  0.32176 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept) -10.461203   0.354234 -29.532  < 2e-16 ***
lnheight      2.220499   0.085019  26.118  < 2e-16 ***
lnage         0.096022   0.032125   2.989  0.00288 ** 
ethnic011    -0.022727   0.009471  -2.400  0.01662 *  
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.1284 on 889 degrees of freedom
  (9 observations deleted due to missingness)
Multiple R-squared:  0.843, Adjusted R-squared:  0.8425 
F-statistic:  1591 on 3 and 889 DF,  p-value: < 2.2e-16
Code 
summary(lm_modfvcmm)

Call:
lm(formula = lnfvc ~ lnheight + lnage + ethnic01, data = mmx)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.45933 -0.08810  0.00582  0.09153  0.41016 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept) -10.990665   0.374745 -29.328  < 2e-16 ***
lnheight      2.329277   0.089966  25.890  < 2e-16 ***
lnage         0.148168   0.033902   4.371 1.38e-05 ***
ethnic011    -0.032887   0.009958  -3.303 0.000995 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.1362 on 898 degrees of freedom
Multiple R-squared:  0.8553,    Adjusted R-squared:  0.8548 
F-statistic:  1769 on 3 and 898 DF,  p-value: < 2.2e-16
Code 
summary(lm_modfev1fvcmm)

Call:
lm(formula = fev1fvc ~ lnheight, data = mm)

Residuals:
      Min        1Q    Median        3Q       Max 
-0.160633 -0.031450  0.000162  0.036037  0.128664 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)  1.80392    0.06619   27.25   <2e-16 ***
lnheight    -0.18563    0.01355  -13.70   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.04988 on 912 degrees of freedom
  (17 observations deleted due to missingness)
Multiple R-squared:  0.1706,    Adjusted R-squared:  0.1697 
F-statistic: 187.6 on 1 and 912 DF,  p-value: < 2.2e-16
Code 
summary(lm_modfefmm)

Call:
lm(formula = lnfef2575 ~ lnheight + lnage, data = mm)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.78772 -0.13297  0.00989  0.14964  0.72185 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept) -6.58183    0.63874 -10.304   <2e-16 ***
lnheight     1.44637    0.15343   9.427   <2e-16 ***
lnage        0.14416    0.05768   2.499   0.0126 *  
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.224 on 822 degrees of freedom
  (106 observations deleted due to missingness)
Multiple R-squared:  0.5049,    Adjusted R-squared:  0.5037 
F-statistic: 419.1 on 2 and 822 DF,  p-value: < 2.2e-16
Code 
summary(lm_modfeffvcmm)

Call:
lm(formula = feffvc ~ lnheight + ethnic01, data = mm)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.67462 -0.17563 -0.02313  0.16388  1.17774 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)  6.61752    0.37321  17.732  < 2e-16 ***
lnheight    -1.12713    0.07675 -14.686  < 2e-16 ***
ethnic011    0.06624    0.02039   3.248  0.00121 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.2694 on 793 degrees of freedom
  (135 observations deleted due to missingness)
Multiple R-squared:  0.2165,    Adjusted R-squared:  0.2145 
F-statistic: 109.5 on 2 and 793 DF,  p-value: < 2.2e-16

Sexo Feminino

Code 
lm_modfev1ff    <- lm(lnfev1    ~ lnheight + lnage + ethnic01, data = ffx)
lm_modfvcff     <- lm(lnfvc     ~ lnheight + lnage + ethnic01, data = ffx)
lm_modfev1fvcff <- lm(fev1fvc   ~ lnheight,                    data = ff)
lm_modfefff     <- lm(lnfef2575 ~ lnheight + lnage,            data = ff)
lm_modfeffvcff  <- lm(feffvc    ~ lnheight + ethnic01,          data = ff)

summary(lm_modfev1ff)

Call:
lm(formula = lnfev1 ~ lnheight + lnage + ethnic01, data = ffx)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.48255 -0.08864 -0.00205  0.08760  0.41210 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept) -10.039170   0.350143 -28.672  < 2e-16 ***
lnheight      2.095265   0.084787  24.712  < 2e-16 ***
lnage         0.169894   0.033672   5.046 5.37e-07 ***
ethnic011    -0.043425   0.009564  -4.540 6.30e-06 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.1369 on 1003 degrees of freedom
  (14 observations deleted due to missingness)
Multiple R-squared:  0.833, Adjusted R-squared:  0.8325 
F-statistic:  1668 on 3 and 1003 DF,  p-value: < 2.2e-16
Code 
summary(lm_modfvcff)

Call:
lm(formula = lnfvc ~ lnheight + lnage + ethnic01, data = ffx)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.41527 -0.09252 -0.00577  0.09298  0.45305 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) -10.57518    0.36720 -28.799  < 2e-16 ***
lnheight      2.21994    0.08882  24.993  < 2e-16 ***
lnage         0.17866    0.03492   5.115 3.74e-07 ***
ethnic011    -0.04767    0.00997  -4.782 1.99e-06 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.1443 on 1017 degrees of freedom
Multiple R-squared:  0.8391,    Adjusted R-squared:  0.8386 
F-statistic:  1768 on 3 and 1017 DF,  p-value: < 2.2e-16
Code 
summary(lm_modfev1fvcff)

Call:
lm(formula = fev1fvc ~ lnheight, data = ff)

Residuals:
      Min        1Q    Median        3Q       Max 
-0.166589 -0.032347  0.003639  0.039452  0.106032 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)  1.42570    0.06265  22.757  < 2e-16 ***
lnheight    -0.10451    0.01282  -8.153 1.02e-15 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.05075 on 1036 degrees of freedom
  (21 observations deleted due to missingness)
Multiple R-squared:  0.06029,   Adjusted R-squared:  0.05939 
F-statistic: 66.47 on 1 and 1036 DF,  p-value: 1.018e-15
Code 
summary(lm_modfefff)

Call:
lm(formula = lnfef2575 ~ lnheight + lnage, data = ff)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.72497 -0.15900  0.00571  0.15514  0.81326 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept) -7.37641    0.61169 -12.059  < 2e-16 ***
lnheight     1.60428    0.14791  10.846  < 2e-16 ***
lnage        0.17286    0.05792   2.984  0.00292 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.2297 on 914 degrees of freedom
  (142 observations deleted due to missingness)
Multiple R-squared:  0.5602,    Adjusted R-squared:  0.5592 
F-statistic: 582.1 on 2 and 914 DF,  p-value: < 2.2e-16
Code 
summary(lm_modfeffvcff)

Call:
lm(formula = feffvc ~ lnheight + ethnic01, data = ff)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.75071 -0.19173 -0.01626  0.16452  1.37370 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)  5.71889    0.37674  15.180   <2e-16 ***
lnheight    -0.92133    0.07746 -11.894   <2e-16 ***
ethnic011    0.05616    0.02126   2.641   0.0084 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.2929 on 876 degrees of freedom
  (180 observations deleted due to missingness)
Multiple R-squared:  0.1401,    Adjusted R-squared:  0.1381 
F-statistic: 71.37 on 2 and 876 DF,  p-value: < 2.2e-16

Estimativa do Efeito da Etnia

Code 
modfev1mm2 <- lm(fev1 ~ lnheight + lnage + ethnic01, data = mmx)
modfvcmm2  <- lm(fvc  ~ lnheight + lnage + ethnic01, data = mmx)

modfev1ff2 <- lm(fev1 ~ lnheight + lnage + ethnic01, data = ffx)
modfvcff2  <- lm(fvc  ~ lnheight + lnage + ethnic01, data = ffx)

summary(modfev1mm2)

Call:
lm(formula = fev1 ~ lnheight + lnage + ethnic01, data = mmx)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.89304 -0.17479 -0.00973  0.15983  1.20708 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept) -18.498405   0.730183 -25.334   <2e-16 ***
lnheight      4.177456   0.175249  23.837   <2e-16 ***
lnage        -0.009394   0.066219  -0.142    0.887    
ethnic011    -0.018437   0.019524  -0.944    0.345    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.2648 on 889 degrees of freedom
  (9 observations deleted due to missingness)
Multiple R-squared:  0.7839,    Adjusted R-squared:  0.7831 
F-statistic:  1075 on 3 and 889 DF,  p-value: < 2.2e-16
Code 
summary(modfvcmm2)

Call:
lm(formula = fvc ~ lnheight + lnage + ethnic01, data = mmx)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.95926 -0.20032 -0.01262  0.16863  1.74798 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) -22.08664    0.88176 -25.048   <2e-16 ***
lnheight      4.94387    0.21169  23.355   <2e-16 ***
lnage         0.03149    0.07977   0.395    0.693    
ethnic011    -0.03320    0.02343  -1.417    0.157    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.3204 on 898 degrees of freedom
Multiple R-squared:  0.7878,    Adjusted R-squared:  0.7871 
F-statistic:  1111 on 3 and 898 DF,  p-value: < 2.2e-16
Code 
summary(modfev1ff2)

Call:
lm(formula = fev1 ~ lnheight + lnage + ethnic01, data = ffx)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.83812 -0.17791 -0.02568  0.16136  1.56884 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) -18.28037    0.74523 -24.530  < 2e-16 ***
lnheight      4.09139    0.18046  22.672  < 2e-16 ***
lnage         0.06166    0.07167   0.860  0.38977    
ethnic011    -0.05844    0.02036  -2.871  0.00418 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.2914 on 1003 degrees of freedom
  (14 observations deleted due to missingness)
Multiple R-squared:  0.7611,    Adjusted R-squared:  0.7604 
F-statistic:  1065 on 3 and 1003 DF,  p-value: < 2.2e-16
Code 
summary(modfvcff2)

Call:
lm(formula = fvc ~ lnheight + lnage + ethnic01, data = ffx)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.90481 -0.20752 -0.02727  0.18180  2.23525 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) -21.01769    0.85158 -24.681  < 2e-16 ***
lnheight      4.69321    0.20599  22.784  < 2e-16 ***
lnage         0.05408    0.08100   0.668  0.50449    
ethnic011    -0.06319    0.02312  -2.733  0.00638 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.3347 on 1017 degrees of freedom
Multiple R-squared:  0.7637,    Adjusted R-squared:  0.763 
F-statistic:  1096 on 3 and 1017 DF,  p-value: < 2.2e-16

Figuras

Histogramas

Code 
ggplot(x2005, aes(age)) +
  geom_histogram(binwidth = 1, fill = "#0072B5", color = "black") +
  facet_wrap(~ sexf) +
  theme_bw() +
  labs(x = "Idade (anos)", y = "Número")
Figure 1: Distribuição de idade por sexo.
Code 
ggplot(x2005, aes(height)) +
  geom_histogram(binwidth = 10, fill = "#0072B5", color = "black") +
  facet_wrap(~ sexf) +
  theme_bw() +
  labs(x = "Estatura (cm)", y = "Número")
Figure 2: Distribuição de estatura por sexo.

Figura 1 — VEF1 e CVF por Etnia (Regressão Mediana + P5)

Code 
log_x_breaks <- c(90, 100, 110, 120, 130, 140, 150, 160, 170, 180)
log_y_breaks <- c(0.5, 0.75, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 5.5, 6)

a1 <- VEF1m +
  geom_point(alpha = 0.2) +
  facet_wrap(~ ethnicf) +
  theme_bw() +
  scale_color_manual(values = palette_nejm) +
  scale_x_log10(name = "Estatura (cm)", breaks = log_x_breaks) +
  scale_y_log10(name = "VEF1 (L)",      breaks = log_y_breaks) +
  theme(legend.position = "none") +
  geom_quantile(quantiles = 0.5, color = "black", linewidth = 1) +
  geom_quantile(quantiles = 0.05, color = "black")

a2 <- CVFm +
  geom_point(alpha = 0.2) +
  facet_wrap(~ ethnicf) +
  theme_bw() +
  scale_color_manual(values = palette_nejm) +
  scale_x_log10(name = "Estatura (cm)", breaks = log_x_breaks) +
  scale_y_log10(name = "CVF (L)",       breaks = log_y_breaks) +
  theme(legend.position = "none") +
  geom_quantile(quantiles = 0.5, color = "black", linewidth = 1) +
  geom_quantile(quantiles = 0.05, color = "black")

a3 <- VEF1f +
  geom_point(alpha = 0.2) +
  facet_wrap(~ ethnicf) +
  theme_bw() +
  scale_color_manual(values = palette_nejm) +
  scale_x_log10(name = "Estatura (cm)", breaks = log_x_breaks) +
  scale_y_log10(name = "VEF1 (L)",      breaks = log_y_breaks) +
  theme(legend.position = "none") +
  geom_quantile(quantiles = 0.5, color = "black", linewidth = 1) +
  geom_quantile(quantiles = 0.05, color = "black")

a4 <- CVFf +
  geom_point(alpha = 0.2) +
  facet_wrap(~ ethnicf) +
  theme_bw() +
  scale_color_manual(values = palette_nejm) +
  scale_x_log10(name = "Estatura (cm)", breaks = log_x_breaks) +
  scale_y_log10(name = "CVF (L)",       breaks = log_y_breaks) +
  theme(legend.position = "none") +
  geom_quantile(quantiles = 0.5, color = "black", linewidth = 1) +
  geom_quantile(quantiles = 0.05, color = "black")

figure1 <- ggarrange(a1, a2, a3, a4, labels = c("A", "B", "C", "D"), ncol = 2, nrow = 2)
figure1
Figure 3: VEF1 e CVF em função da estatura, estratificados por etnia. Masculino (A, B) e Feminino (C, D). Linha grossa = mediana (P50); linha fina = limite inferior da normalidade (P5).

Figura 2 — VEF1/CVF e FEF25-75 por Sexo (Regressão Mediana + P5)

Code 
b1 <- TIFFENAU +
  geom_point(alpha = 0.2) +
  facet_wrap(~ sexf) +
  theme_bw() +
  scale_x_continuous(name = "Estatura (cm)", breaks = c(90, 100, 110, 120, 130, 140, 150, 160, 170, 180)) +
  scale_y_continuous(name = "VEF1/CVF",      breaks = c(0.7, 0.75, 0.8, 0.85, 0.9, 0.95, 1)) +
  theme(legend.position = "none") +
  geom_quantile(quantiles = 0.5,  color = "black", linewidth = 1) +
  geom_quantile(quantiles = 0.05, color = "black")

b2 <- FEF2575 +
  geom_point(alpha = 0.2) +
  facet_wrap(~ sexf) +
  theme_bw() +
  scale_x_log10(name = "Estatura (cm)",    breaks = c(90, 100, 110, 120, 130, 140, 150, 160, 170, 180)) +
  scale_y_log10(name = "FEF25-75% (L/s)", breaks = c(0.5, 0.75, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 5.5, 6)) +
  theme(legend.position = "none") +
  geom_quantile(quantiles = 0.5,  color = "black", linewidth = 1) +
  geom_quantile(quantiles = 0.05, color = "black")

figure2 <- ggarrange(b1, b2, labels = c("A", "B"), ncol = 1, nrow = 2)
figure2
Figure 4: VEF1/CVF (A) e FEF25-75 (B) por sexo. Linha grossa = mediana (P50); linha fina = limite inferior da normalidade (P5).

Figura 3 — Diferença Percentual em Relação ao GLI2012 e Pereira por Etnia

Code 
c1 <- ggplot(x2005x, aes(ethnicnpb, gpvef1, fill = ethnicnpb)) +
  geom_boxplot(outlier.alpha = 0) +
  labs(x = "", y = "Diferença percentual no VEF1") +
  scale_fill_brewer(palette = "Blues") +
  theme_bw() +
  theme(legend.position = "none") +
  ylim(-35, 60)

c2 <- ggplot(x2005x, aes(ethnicnpb, gpcvf, fill = ethnicnpb)) +
  geom_boxplot(outlier.alpha = 0) +
  labs(x = "", y = "Diferença percentual na CVF") +
  scale_fill_brewer(palette = "Blues") +
  theme_bw() +
  theme(legend.position = "none") +
  ylim(-35, 60)

c3 <- ggplot(x2005x, aes(ethnicnpb, ppvef1, fill = ethnicnpb)) +
  geom_boxplot(outlier.alpha = 0) +
  labs(x = "", y = "Diferença percentual no VEF1") +
  scale_fill_brewer(palette = "Greens") +
  theme_bw() +
  theme(legend.position = "none") +
  ylim(-35, 60)

c4 <- ggplot(x2005x, aes(ethnicnpb, ppcvf, fill = ethnicnpb)) +
  geom_boxplot(outlier.alpha = 0) +
  labs(x = "", y = "Diferença percentual na CVF") +
  scale_fill_brewer(palette = "Greens") +
  theme_bw() +
  theme(legend.position = "none") +
  ylim(-35, 60)

figure3 <- ggarrange(c1, c2, c3, c4, labels = c("A", "B", "C", "D"))
figure3
Figure 5: Diferença percentual entre valores observados e preditos pelo GLI2012 (A, B) e Pereira (C, D), por grupo étnico.

Figura 3a — Diferença Percentual por Etnia e Sexo

Code 
c1 <- ggplot(x2005x, aes(ethnicnpb, gpvef1, fill = ethnicnpb)) +
  geom_boxplot(outlier.alpha = 0) +
  facet_wrap(~ sexf) +
  labs(x = "", y = "Diferença percentual no VEF1") +
  scale_fill_brewer(palette = "Blues") +
  theme_bw() +
  theme(legend.position = "none")

c2 <- ggplot(x2005x, aes(ethnicnpb, gpcvf, fill = ethnicnpb)) +
  geom_boxplot(outlier.alpha = 0) +
  facet_wrap(~ sexf) +
  labs(x = "", y = "Diferença percentual na CVF") +
  scale_fill_brewer(palette = "Blues") +
  theme_bw() +
  theme(legend.position = "none")

c3 <- ggplot(x2005x, aes(ethnicnpb, ppvef1, fill = ethnicnpb)) +
  geom_boxplot(outlier.alpha = 0) +
  facet_wrap(~ sexf) +
  labs(x = "", y = "Diferença percentual no VEF1") +
  scale_fill_brewer(palette = "Greens") +
  theme_bw() +
  theme(legend.position = "none")

c4 <- ggplot(x2005x, aes(ethnicnpb, ppcvf, fill = ethnicnpb)) +
  geom_boxplot(outlier.alpha = 0) +
  facet_wrap(~ sexf) +
  labs(x = "", y = "Diferença percentual na CVF") +
  scale_fill_brewer(palette = "Greens") +
  theme_bw() +
  theme(legend.position = "none")

figure3a <- ggarrange(c1, c2, c3, c4, labels = c("A", "B", "C", "D"))
figure3a
Figure 6: Diferença percentual em relação ao GLI2012 (A, B) e Pereira (C, D), estratificada por etnia e sexo.

Tabela 2 — Comparação com GLI2012 e Pereira

Z-Scores GLI2012

Code 
expss_digits(digits = 3)
x2005 |> calc_cro_mean_sd_n(fev1z,    ethnicnpb, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(fvcz,     ethnicnpb, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(fev1fvcz, ethnicnpb, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(fef2575z, ethnicnpb, sexf) |> htmlTable()
Table 2
   ethnicnpb 
   Branca   Negra   Parda 
 sexf 
   Feminino   fev1z   Mean    0.353 1.158 0.604
    Std. dev.    1.159 1.267 1.290
    Unw. valid N    715.000 101.000 191.000
   Masculino   fev1z   Mean    0.293 1.317 0.746
    Std. dev.    1.080 1.182 1.279
    Unw. valid N    627.000 80.000 186.000
   ethnicnpb 
   Branca   Negra   Parda 
 sexf 
   Feminino   fvcz   Mean    0.159 0.987 0.485
    Std. dev.    1.178 1.254 1.352
    Unw. valid N    722.000 102.000 197.000
   Masculino   fvcz   Mean    0.098 1.131 0.600
    Std. dev.    1.094 1.224 1.373
    Unw. valid N    631.000 82.000 189.000
   ethnicnpb 
   Branca   Negra   Parda 
 sexf 
   Feminino   fev1fvcz   Mean    0.334 0.248 0.218
    Std. dev.    0.996 0.997 1.053
    Unw. valid N    715.000 101.000 191.000
   Masculino   fev1fvcz   Mean    0.322 0.254 0.253
    Std. dev.    0.914 0.902 0.973
    Unw. valid N    627.000 80.000 186.000
   ethnicnpb 
   Branca   Negra   Parda 
 sexf 
   Feminino   fef2575z   Mean    0.232 0.505 0.374
    Std. dev.    1.040 0.884 1.069
    Unw. valid N    593.000 95.000 191.000
   Masculino   fef2575z   Mean    0.185 0.707 0.463
    Std. dev.    0.934 0.824 1.056
    Unw. valid N    533.000 80.000 183.000

Diferença Percentual em Relação ao GLI2012

Code 
expss_digits(digits = 1)
x2005 |> calc_cro_mean_sd_n(gpvef1,    ethnicnpb, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(gpcvf,     ethnicnpb, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(gpfef2575, ethnicnpb, sexf) |> htmlTable()
Table 3
   ethnicnpb 
   Branca   Negra   Parda 
 sexf 
   Feminino   gpvef1   Mean    4.1 14.9 7.1
    Std. dev.    13.8 16.3 15.3
    Unw. valid N    715.0 101.0 191.0
   Masculino   gpvef1   Mean    3.5 17.3 8.7
    Std. dev.    12.8 15.6 15.0
    Unw. valid N    627.0 80.0 186.0
   ethnicnpb 
   Branca   Negra   Parda 
 sexf 
   Feminino   gpcvf   Mean    2.2 13.3 5.8
    Std. dev.    14.6 16.9 16.0
    Unw. valid N    722.0 102.0 197.0
   Masculino   gpcvf   Mean    1.3 15.0 6.8
    Std. dev.    13.3 16.5 15.5
    Unw. valid N    631.0 82.0 189.0
   ethnicnpb 
   Branca   Negra   Parda 
 sexf 
   Feminino   gpfef2575   Mean    5.7 14.4 9.2
    Std. dev.    24.2 24.5 24.8
    Unw. valid N    593.0 95.0 191.0
   Masculino   gpfef2575   Mean    6.0 22.4 12.7
    Std. dev.    23.6 25.3 27.0
    Unw. valid N    533.0 80.0 182.0

Diferença Absoluta em Relação ao GLI2012

Code 
expss_digits(digits = 3)
x2005 |> calc_cro_mean_sd_n(dvef1g,    ethnic, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(dcvfg,     ethnic, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(dvef1cvfg, ethnic, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(dfmefg,    ethnic, sexf) |> htmlTable()
Table 4
   ethnic 
   1   2   5 
 sexf 
   Feminino   dvef1g   Mean    0.066 0.249 0.124
    Std. dev.    0.252 0.304 0.286
    Unw. valid N    715.000 101.000 191.000
   Masculino   dvef1g   Mean    0.055 0.262 0.159
    Std. dev.    0.240 0.258 0.293
    Unw. valid N    627.000 80.000 186.000
   ethnic 
   1   2   5 
 sexf 
   Feminino   dcvfg   Mean    0.034 0.251 0.114
    Std. dev.    0.299 0.352 0.325
    Unw. valid N    722.000 102.000 197.000
   Masculino   dcvfg   Mean    0.017 0.257 0.147
    Std. dev.    0.291 0.296 0.351
    Unw. valid N    631.000 82.000 189.000
   ethnic 
   1   2   5 
 sexf 
   Feminino   dvef1cvfg   Mean    0.013 0.008 0.007
    Std. dev.    0.050 0.052 0.050
    Unw. valid N    711.000 101.000 189.000
   Masculino   dvef1cvfg   Mean    0.017 0.011 0.012
    Std. dev.    0.050 0.051 0.050
    Unw. valid N    622.000 80.000 184.000
   ethnic 
   1   2   5 
 sexf 
   Feminino   dfmefg   Mean    0.130 0.313 0.192
    Std. dev.    0.551 0.553 0.570
    Unw. valid N    593.000 95.000 191.000
   Masculino   dfmefg   Mean    0.114 0.419 0.290
    Std. dev.    0.501 0.495 0.695
    Unw. valid N    533.000 80.000 183.000

Diferença Percentual em Relação a Pereira

Code 
expss_digits(digits = 1)
x2005 |> calc_cro_mean_sd_n(ppvef1, ethnic, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(ppcvf,  ethnic, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(ppfmef, ethnic, sexf) |> htmlTable()
Table 5
   ethnic 
   1   2   5 
 sexf 
   Feminino   ppvef1   Mean    9.7 3.6 5.8
    Std. dev.    15.1 15.3 14.2
    Unw. valid N    574.0 81.0 168.0
   Masculino   ppvef1   Mean    5.6 1.9 4.8
    Std. dev.    13.1 13.6 14.4
    Unw. valid N    492.0 62.0 165.0
   ethnic 
   1   2   5 
 sexf 
   Feminino   ppcvf   Mean    12.5 6.9 9.1
    Std. dev.    16.2 16.0 15.4
    Unw. valid N    575.0 81.0 168.0
   Masculino   ppcvf   Mean    10.2 6.1 8.9
    Std. dev.    14.4 14.3 15.5
    Unw. valid N    492.0 62.0 165.0
   ethnic 
   1   2   5 
 sexf 
   Feminino   ppfmef   Mean    13.7 7.3 8.5
    Std. dev.    25.7 23.9 25.8
    Unw. valid N    447.0 74.0 162.0
   Masculino   ppfmef   Mean    1.3 2.5 3.8
    Std. dev.    23.1 21.4 25.2
    Unw. valid N    395.0 60.0 159.0

Diferença Absoluta em Relação a Pereira

Code 
expss_digits(digits = 3)
x2005 |> calc_cro_mean_sd_n(dvef1p,    ethnic, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(dcvfp,     ethnic, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(dvef1cvfp, ethnic, sexf) |> htmlTable()
x2005 |> calc_cro_mean_sd_n(dfmefp,    ethnic, sexf) |> htmlTable()
Table 6
   ethnic 
   1   2   5 
 sexf 
   Feminino   dvef1p   Mean    0.156 0.059 0.109
    Std. dev.    0.297 0.325 0.283
    Unw. valid N    575.000 81.000 168.000
   Masculino   dvef1p   Mean    0.101 0.043 0.096
    Std. dev.    0.262 0.276 0.296
    Unw. valid N    492.000 62.000 165.000
   ethnic 
   1   2   5 
 sexf 
   Feminino   dcvfp   Mean    0.229 0.134 0.181
    Std. dev.    0.319 0.371 0.329
    Unw. valid N    575.000 81.000 168.000
   Masculino   dcvfp   Mean    0.205 0.132 0.192
    Std. dev.    0.319 0.318 0.364
    Unw. valid N    492.000 62.000 165.000
   ethnic 
   1   2   5 
 sexf 
   Feminino   dvef1cvfp   Mean    -0.019 -0.030 -0.022
    Std. dev.    0.050 0.054 0.052
    Unw. valid N    570.000 81.000 166.000
   Masculino   dvef1cvfp   Mean    -0.042 -0.043 -0.038
    Std. dev.    0.052 0.052 0.053
    Unw. valid N    487.000 62.000 163.000
   ethnic 
   1   2   5 
 sexf 
   Feminino   dfmefp   Mean    0.293 0.165 0.185
    Std. dev.    0.592 0.598 0.594
    Unw. valid N    447.000 74.000 162.000
   Masculino   dfmefp   Mean    0.017 0.068 0.082
    Std. dev.    0.548 0.525 0.616
    Unw. valid N    395.000 60.000 159.000

Testes de Wilcoxon: Observado vs. Pereira

Code 
# Helper para simplificar chamadas repetidas
wx_paired <- function(obs, pred) {
  wilcox.test(obs, pred, paired = TRUE)
}

ethnicv <- c(1, 2, 5)

for (sx in 1:2) {
  sex_label <- if (sx == 1) "MASCULINO" else "FEMININO"
  cat("\n====", sex_label, "====\n")
  for (e in ethnicv) {
    idx <- x2005$sex == sx & x2005$ethnic == e
    cat("\nEtnia:", e, "\n")
    cat("VEF1:    "); print(wx_paired(x2005$fev1[idx],    x2005$vef1prev[idx]))
    cat("CVF:     "); print(wx_paired(x2005$fvc[idx],     x2005$cvfprev[idx]))
    cat("VEF1/CVF:"); print(wx_paired(x2005$fev1fvc[idx], x2005$vef1cvfprev[idx]))
    cat("FEF25-75:"); print(wx_paired(x2005$fef2575[idx], x2005$fmefprev[idx]))
  }
}

==== MASCULINO ====

Etnia: 1 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 86783, p-value < 2.2e-16
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 101470, p-value < 2.2e-16
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 15308, p-value < 2.2e-16
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 39456, p-value = 0.8773
alternative hypothesis: true location shift is not equal to 0


Etnia: 2 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 1147, p-value = 0.2333
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 1433, p-value = 0.001388
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 230, p-value = 1.693e-07
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 1041, p-value = 0.3555
alternative hypothesis: true location shift is not equal to 0


Etnia: 5 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 9166, p-value = 0.0001623
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 10665, p-value = 5.286e-10
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 2165, p-value = 7.134e-14
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 6939, p-value = 0.3198
alternative hypothesis: true location shift is not equal to 0


==== FEMININO ====

Etnia: 1 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 132525, p-value < 2.2e-16
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 142212, p-value < 2.2e-16
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 50587, p-value = 5.075e-15
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 76445, p-value < 2.2e-16
alternative hypothesis: true location shift is not equal to 0


Etnia: 2 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 2032, p-value = 0.08068
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 2354, p-value = 0.001103
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 759, p-value = 2.214e-05
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 1851, p-value = 0.01262
alternative hypothesis: true location shift is not equal to 0


Etnia: 5 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 9959, p-value = 5.888e-06
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 11270, p-value = 3.93e-11
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 3993, p-value = 2.184e-06
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 8639, p-value = 0.0006581
alternative hypothesis: true location shift is not equal to 0

Testes de Wilcoxon: Observado vs. GLI2012

Code 
for (sx in 1:2) {
  sex_label <- if (sx == 1) "MASCULINO" else "FEMININO"
  cat("\n====", sex_label, "====\n")
  for (e in ethnicv) {
    idx <- x2005$sex == sx & x2005$ethnic == e
    cat("\nEtnia:", e, "\n")
    cat("VEF1:    "); print(wx_paired(x2005$fev1[idx],    x2005$fev1gli[idx]))
    cat("CVF:     "); print(wx_paired(x2005$fvc[idx],     x2005$fvcgli[idx]))
    cat("VEF1/CVF:"); print(wx_paired(x2005$fev1fvc[idx], x2005$fev1fvcgli[idx]))
    cat("FEF25-75:"); print(wx_paired(x2005$fef2575[idx], x2005$fef2575gli[idx]))
  }
}

==== MASCULINO ====

Etnia: 1 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 123813, p-value = 2.247e-08
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 107576, p-value = 0.08551
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 132376, p-value = 2.42e-15
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 87971, p-value = 2.279e-06
alternative hypothesis: true location shift is not equal to 0


Etnia: 2 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 3021, p-value = 1.852e-11
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 3060, p-value = 3.431e-10
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 2034, p-value = 0.04734
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 2879, p-value = 1.578e-09
alternative hypothesis: true location shift is not equal to 0


Etnia: 5 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 13617, p-value = 2.185e-11
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 12897, p-value = 1.948e-07
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 11063, p-value = 0.0004183
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 12157, p-value = 7.39e-08
alternative hypothesis: true location shift is not equal to 0


==== FEMININO ====

Etnia: 1 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 164695, p-value = 3.045e-11
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 144568, p-value = 0.01211
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 166525, p-value = 2.987e-13
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 107415, p-value = 3.536e-06
alternative hypothesis: true location shift is not equal to 0


Etnia: 2 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 4606, p-value = 6.117e-12
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 4545, p-value = 1.527e-10
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 3096, p-value = 0.07814
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 3601, p-value = 9.51e-07
alternative hypothesis: true location shift is not equal to 0


Etnia: 5 
VEF1:    
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 13439, p-value = 2.373e-08
alternative hypothesis: true location shift is not equal to 0

CVF:     
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 13105, p-value = 2.854e-05
alternative hypothesis: true location shift is not equal to 0

VEF1/CVF:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 10788, p-value = 0.01624
alternative hypothesis: true location shift is not equal to 0

FEF25-75:
    Wilcoxon signed rank test with continuity correction

data:  obs and pred
V = 12286, p-value = 4.598e-05
alternative hypothesis: true location shift is not equal to 0

Comparação com GLI Caucasiano Aplicado a Todos

Code 
X2005_GLI_WHITE <- read_excel("2005 GLI WHITE.xlsx")
gw <- X2005_GLI_WHITE
names(gw) <- tolower(names(gw))

# Rename slash-separated column names to safe R names
gw <- gw |>
  rename(
    fev1fvcz   = `fev1/fvcz`,
    fev1fvclln = `fev1/fvclln`,
    fev1fvcpred = `fev1/fvcpred`,
    fev1fvctile = `fev1/fvctile`,
    fev1fvc    = `fev1/fvc`
  ) |>
  mutate(
    sexf     = as.factor(sex),
    ethnic   = as.factor(ethnic),
    ethnic12 = as.factor(ethnic12),
    ethnicw  = as.factor(ethnicw)
  )

expss_digits(digits = 3)
gw |> calc_cro_mean_sd_n(fev1z,    ethnic, sexf) |> htmlTable()
   ethnic 
   1   2   5 
 sexf 
   Female   fev1z   Mean    0.351 -0.061 -0.024
    Std. dev.    1.166 1.174 1.220
    Unw. valid N    718.000 101.000 192.000
   Male   fev1z   Mean    0.289 0.009 0.109
    Std. dev.    1.101 1.113 1.198
    Unw. valid N    634.000 80.000 188.000
Code 
gw |> calc_cro_mean_sd_n(fvcz,     ethnic, sexf) |> htmlTable()
   ethnic 
   1   2   5 
 sexf 
   Female   fvcz   Mean    0.160 -0.247 -0.240
    Std. dev.    1.186 1.184 1.209
    Unw. valid N    725.000 102.000 198.000
   Male   fvcz   Mean    0.106 -0.236 -0.133
    Std. dev.    1.113 1.130 1.200
    Unw. valid N    638.000 82.000 191.000
Code 
gw |> calc_cro_mean_sd_n(fev1fvcz, ethnic, sexf) |> htmlTable()
   ethnic 
   1   2   5 
 sexf 
   Female   fev1fvcz   Mean    0.328 0.340 0.383
    Std. dev.    0.999 0.990 1.018
    Unw. valid N    718.000 101.000 192.000
   Male   fev1fvcz   Mean    0.304 0.372 0.380
    Std. dev.    0.929 0.891 0.941
    Unw. valid N    634.000 80.000 188.000
Code 
gw |> calc_cro_mean_sd_n(fef2575z, ethnic, sexf) |> htmlTable()
   ethnic 
   1   2   5 
 sexf 
   Female   fef2575z   Mean    0.221 0.029 0.086
    Std. dev.    1.060 0.960 1.082
    Unw. valid N    598.000 95.000 194.000
   Male   fef2575z   Mean    0.176 0.228 0.198
    Std. dev.    0.951 0.899 1.071
    Unw. valid N    540.000 80.000 186.000

Matrizes de Correlação

Code 
vars <- c("age", "height", "weight", "imc", "fev1", "fvc", "fef2575")

cor_mm <- mm |> select(all_of(vars)) |> na.omit() |> cor()
cor_ff <- ff |> select(all_of(vars)) |> na.omit() |> cor()

p_mm <- ggcorrplot(
  cor_mm, hc.order = FALSE, type = "lower", lab = TRUE, lab_size = 3,
  method = "circle", colors = c("tomato2", "white", "springgreen3"),
  title = "Masculino", ggtheme = theme_bw
)

p_ff <- ggcorrplot(
  cor_ff, hc.order = FALSE, type = "lower", lab = TRUE, lab_size = 3,
  method = "circle", colors = c("tomato2", "white", "springgreen3"),
  title = "Feminino", ggtheme = theme_bw
)

p_mm + p_ff
Figure 7: Correlograma das variáveis antropométricas e espirométricas por sexo.

Modelos GAMLSS

Preparação dos Dados

Code 
vars_base <- c("height", "age", "ethnic01")

mm_clean_fvc     <- na.omit(mm[, c("fvc",     vars_base)])
mm_clean_fev1    <- na.omit(mm[, c("fev1",    vars_base)])
mm_clean_fev1fvc <- na.omit(mm[, c("fev1fvc", vars_base)])
mm_clean_fef2575 <- na.omit(mm[, c("fef2575", vars_base)])

ff_clean_fvc     <- na.omit(ff[, c("fvc",     vars_base)])
ff_clean_fev1    <- na.omit(ff[, c("fev1",    vars_base)])
ff_clean_fev1fvc <- na.omit(ff[, c("fev1fvc", vars_base)])
ff_clean_fef2575 <- na.omit(ff[, c("fef2575", vars_base)])

Modelos Masculinos — Escala Original

Code 
gamlss_formula <- function(outcome, data) {
  gamlss(
    as.formula(paste(outcome, "~ pb(height) + pb(age) + ethnic01")),
    sigma.formula = ~ pb(height) + pb(age) + ethnic01,
    family  = BCCGo,
    data    = data,
    trace   = FALSE
  )
}

mm_model_fvc     <- gamlss_formula("fvc",     mm_clean_fvc)
mm_model_fev1    <- gamlss_formula("fev1",    mm_clean_fev1)
mm_model_fev1fvc <- gamlss_formula("fev1fvc", mm_clean_fev1fvc)
mm_model_fef2575 <- gamlss_formula("fef2575", mm_clean_fef2575)

Modelos Femininos

Code 
ff_model_fvc     <- gamlss_formula("fvc",     ff_clean_fvc)
ff_model_fev1    <- gamlss_formula("fev1",    ff_clean_fev1)
ff_model_fev1fvc <- gamlss_formula("fev1fvc", ff_clean_fev1fvc)
ff_model_fef2575 <- gamlss_formula("fef2575", ff_clean_fef2575)

Modelos GAMLSS com Pesos por Etnia

Code 
spirometry_data <- x2005 |>
  mutate(weight_ethnicity = case_when(
    ethnic == 1 ~ 0.64,
    ethnic == 2 ~ 2.38,
    ethnic == 5 ~ 1.13
  ))

data_female <- spirometry_data |>
  filter(sex == 2) |>
  select(fev1, age, height, weight_ethnicity) |>
  na.omit()

fev1_model_female <- gamlss(
  fev1 ~ pb(height) + pb(age),
  sigma.formula = ~ pb(height) + pb(age),
  nu.formula    = ~ 1,
  tau.formula   = ~ 1,
  family        = BCCGo,
  weights       = weight_ethnicity,
  data          = data_female,
  trace         = FALSE
)

summary(fev1_model_female)
******************************************************************
Family:  c("BCCGo", "Box-Cox-Cole-Green-orig.") 

Call:  
gamlss(formula = fev1 ~ pb(height) + pb(age), sigma.formula = ~pb(height) +  
    pb(age), nu.formula = ~1, tau.formula = ~1, family = BCCGo,  
    data = data_female, weights = weight_ethnicity,      trace = FALSE) 

Fitting method: RS() 

------------------------------------------------------------------
Mu link function:  log
Mu Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept) -1.7870037  0.0566655 -31.536  < 2e-16 ***
pb(height)   0.0157139  0.0006729  23.351  < 2e-16 ***
pb(age)      0.0262923  0.0044828   5.865 6.11e-09 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

------------------------------------------------------------------
Sigma link function:  log
Sigma Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) -2.509819   0.263201  -9.536   <2e-16 ***
pb(height)   0.005824   0.003129   1.861    0.063 .  
pb(age)     -0.029411   0.021806  -1.349    0.178    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

------------------------------------------------------------------
Nu link function:  identity 
Nu Coefficients:
            Estimate Std. Error t value Pr(>|t|)  
(Intercept)   0.3472     0.1961    1.77   0.0771 .
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

------------------------------------------------------------------
NOTE: Additive smoothing terms exist in the formulas: 
 i) Std. Error for smoothers are for the linear effect only. 
ii) Std. Error for the linear terms maybe are not accurate. 
------------------------------------------------------------------
No. of observations in the fit:  1007 
Degrees of Freedom for the fit:  13.53572
      Residual Deg. of Freedom:  993.4643 
                      at cycle:  8 
 
Global Deviance:     -40.22937 
            AIC:     -13.15794 
            SBC:     53.36646 
******************************************************************

Predição — Exemplo (10 anos, 140 cm, feminino)

Code 
newdata <- data.frame(age = 10, height = 140)

mu    <- predict(fev1_model_female, newdata = newdata, what = "mu",    type = "response")
sigma <- predict(fev1_model_female, newdata = newdata, what = "sigma", type = "response")
nu    <- predict(fev1_model_female, newdata = newdata, what = "nu",    type = "response")

fev1_median <- mu
fev1_p5     <- qBCCGo(0.05, mu = mu, sigma = sigma, nu = nu)

cat("VEF1 previsto (mediana):", round(fev1_median, 2), "L\n")
VEF1 previsto (mediana): 1.95 L
Code 
cat("Limite inferior da normalidade (P5):", round(fev1_p5, 2), "L\n")
Limite inferior da normalidade (P5): 1.55 L

Comparação dos Limites Inferiores — Nova Equação vs. GLI2012

Carregamento dos Dados

Code 
comp <- read_excel("compare_2020.xlsx", sheet = "raw")
names(comp) <- tolower(names(comp))

# Renomear colunas com hífen para nomes válidos em R
comp <- comp |>
  rename(
    fef2575_p5  = `fef25-75_p5`,
    fef2575_lin = `fef25-75lin`
  ) |>
  mutate(
    agey    = as.factor(round(age, digits = 0)),
    agex    = case_when(
      age <  6             ~ "3-6 anos",
      age >= 6 & age < 9  ~ "6-9 anos",
      age >= 9            ~ "9-12 anos"
    ) |> as.factor(),
    ethnic.f = case_when(
      ethnic == 0 ~ "White",
      ethnic == 1 ~ "Non-White"
    )
  )

comp_f0 <- comp |> filter(sex == "Female", ethnic == 0)
comp_f1 <- comp |> filter(sex == "Female", ethnic == 1)
comp_m0 <- comp |> filter(sex == "Male",   ethnic == 0)
comp_m1 <- comp |> filter(sex == "Male",   ethnic == 1)

Estatísticas Comparativas do Limite Inferior

Code 
expss_digits(digits = 2)

# FVC
cat("== FVC ==\n")
== FVC ==
Code 
calc_cro_mean(comp_m0, (cvf_p5 / fvclln - 1) * 100)
calc_cro_mean(comp_m1, (cvf_p5 / fvclln - 1) * 100)
calc_cro_mean(comp_f0, (cvf_p5 / fvclln - 1) * 100)
calc_cro_mean(comp_f1, (cvf_p5 / fvclln - 1) * 100)
# FEV1
cat("== VEF1 ==\n")
== VEF1 ==
Code 
calc_cro_mean(comp_m0, (vef1_p5 / fev1lln - 1) * 100)
calc_cro_mean(comp_m1, (vef1_p5 / fev1lln - 1) * 100)
calc_cro_mean(comp_f0, (vef1_p5 / fev1lln - 1) * 100)
calc_cro_mean(comp_f1, (vef1_p5 / fev1lln - 1) * 100)
# FEV1/FVC
expss_digits(digits = 3)
cat("== VEF1/CVF ==\n")
== VEF1/CVF ==
Code 
calc_cro_mean(comp_m0, vef1cvf_p5 - fev1fvclln)
calc_cro_mean(comp_m1, vef1cvf_p5 - fev1fvclln)
calc_cro_mean(comp_f0, vef1cvf_p5 - fev1fvclln)
calc_cro_mean(comp_f1, vef1cvf_p5 - fev1fvclln)
# FEF25-75
expss_digits(digits = 2)
cat("== FEF25-75 ==\n")
== FEF25-75 ==
Code 
calc_cro_mean(comp_m0, (fef2575_p5 / fef2575lln - 1) * 100)
calc_cro_mean(comp_m1, (fef2575_p5 / fef2575lln - 1) * 100)
calc_cro_mean(comp_f0, (fef2575_p5 / fef2575lln - 1) * 100)
calc_cro_mean(comp_f1, (fef2575_p5 / fef2575lln - 1) * 100)
Table 7
 #Total 
 (cvf_p5/fvclln - 1) * 100 
 #Total 
 (cvf_p5/fvclln - 1) * 100  0.36
 #Total 
 (cvf_p5/fvclln - 1) * 100 
 #Total 
 (cvf_p5/fvclln - 1) * 100  0.14
 #Total 
 (vef1_p5/fev1lln - 1) * 100 
 #Total 
 (vef1_p5/fev1lln - 1) * 100  3.48
 #Total 
 (vef1_p5/fev1lln - 1) * 100 
 #Total 
 (vef1_p5/fev1lln - 1) * 100  2.09
 #Total 
 vef1cvf_p5 - fev1fvclln 
 #Total 
 vef1cvf_p5 - fev1fvclln  0.045
 #Total 
 vef1cvf_p5 - fev1fvclln 
 #Total 
 vef1cvf_p5 - fev1fvclln  0.028
 #Total 
 (fef2575_p5/fef2575lln - 1) * 100 
 #Total 
 (fef2575_p5/fef2575lln - 1) * 100  10.62
 #Total 
 (fef2575_p5/fef2575lln - 1) * 100 
 #Total 
 (fef2575_p5/fef2575lln - 1) * 100  10.22

Figuras Suplementares — Limite Inferior por Grupo

Code 
# Função auxiliar para construir os painéis de LLN
lln_panel <- function(data, y_obs, y_gli, y_label) {
  ggplot(data) +
    geom_point(aes(height, .data[[y_obs]]), col = "#BC3C29FF", alpha = 0.1) +
    geom_point(aes(height, .data[[y_gli]]), col = "#0072B5FF", alpha = 0.1) +
    geom_smooth(aes(height, .data[[y_obs]]), col = "#BC3C29FF", se = FALSE) +
    geom_smooth(aes(height, .data[[y_gli]]), col = "#0072B5FF", se = FALSE) +
    labs(x = "Estatura (cm)", y = y_label) +
    theme_bw()
}
Code 
s3a <- lln_panel(comp_f0, "vef1_p5",    "fev1lln",    "VEF1 (L) LIN")
s3b <- lln_panel(comp_f0, "cvf_p5",     "fvclln",     "CVF (L) LIN")
s3c <- ggplot(comp_f0) +
  geom_point(aes(height, vef1cvf_p5),  col = "#BC3C29FF", alpha = 0.1) +
  geom_point(aes(height, fev1fvclln),  col = "#0072B5FF", alpha = 0.1) +
  geom_smooth(aes(height, vef1cvf_p5), col = "#BC3C29FF", se = FALSE) +
  geom_smooth(aes(height, fev1fvclln), col = "#0072B5FF", se = FALSE) +
  labs(x = "Estatura (cm)") +
  scale_y_continuous(name = "VEF1/CVF LIN", limits = c(0.7, 0.95)) +
  theme_bw()
s3d <- lln_panel(comp_f0, "fef2575_p5", "fef2575lln", "FEF25-75 (L/s) LIN")
figure_s3 <- ggarrange(s3a, s3b, s3c, s3d, labels = c("A", "B", "C", "D"))
figure_s3
Figure 8: Limite inferior da normalidade: nova equação (vermelho) vs. GLI2012 (azul) — Feminino Branca.
Code 
s4a <- lln_panel(comp_f1, "vef1_p5",    "fev1lln",    "VEF1 (L) LIN")
s4b <- lln_panel(comp_f1, "cvf_p5",     "fvclln",     "CVF (L) LIN")
s4c <- ggplot(comp_f1) +
  geom_point(aes(height, vef1cvf_p5),  col = "#BC3C29FF", alpha = 0.1) +
  geom_point(aes(height, fev1fvclln),  col = "#0072B5FF", alpha = 0.1) +
  geom_smooth(aes(height, vef1cvf_p5), col = "#BC3C29FF", se = FALSE) +
  geom_smooth(aes(height, fev1fvclln), col = "#0072B5FF", se = FALSE) +
  labs(x = "Estatura (cm)") +
  scale_y_continuous(name = "VEF1/CVF LIN", limits = c(0.7, 0.95)) +
  theme_bw()
s4d <- lln_panel(comp_f1, "fef2575_p5", "fef2575lln", "FEF25-75 (L/s) LIN")
figure_s4 <- ggarrange(s4a, s4b, s4c, s4d, labels = c("A", "B", "C", "D"))
figure_s4
Figure 9: Limite inferior da normalidade: nova equação (vermelho) vs. GLI2012 (azul) — Feminino Negra/Parda.
Code 
s5a <- lln_panel(comp_m0, "vef1_p5",    "fev1lln",    "VEF1 (L) LIN")
s5b <- lln_panel(comp_m0, "cvf_p5",     "fvclln",     "CVF (L) LIN")
s5c <- ggplot(comp_m0) +
  geom_point(aes(height, vef1cvf_p5),  col = "#BC3C29FF", alpha = 0.1) +
  geom_point(aes(height, fev1fvclln),  col = "#0072B5FF", alpha = 0.1) +
  geom_smooth(aes(height, vef1cvf_p5), col = "#BC3C29FF", se = FALSE) +
  geom_smooth(aes(height, fev1fvclln), col = "#0072B5FF", se = FALSE) +
  labs(x = "Estatura (cm)") +
  scale_y_continuous(name = "VEF1/CVF LIN", limits = c(0.7, 0.95)) +
  theme_bw()
s5d <- lln_panel(comp_m0, "fef2575_p5", "fef2575lln", "FEF25-75 (L/s) LIN")
figure_s5 <- ggarrange(s5a, s5b, s5c, s5d, labels = c("A", "B", "C", "D"))
figure_s5
Figure 10: Limite inferior da normalidade: nova equação (vermelho) vs. GLI2012 (azul) — Masculino Branco.
Code 
s6a <- lln_panel(comp_m1, "vef1_p5",    "fev1lln",    "VEF1 (L) LIN")
s6b <- lln_panel(comp_m1, "cvf_p5",     "fvclln",     "CVF (L) LIN")
s6c <- ggplot(comp_m1) +
  geom_point(aes(height, vef1cvf_p5),  col = "#BC3C29FF", alpha = 0.1) +
  geom_point(aes(height, fev1fvclln),  col = "#0072B5FF", alpha = 0.1) +
  geom_smooth(aes(height, vef1cvf_p5), col = "#BC3C29FF", se = FALSE) +
  geom_smooth(aes(height, fev1fvclln), col = "#0072B5FF", se = FALSE) +
  labs(x = "Estatura (cm)") +
  scale_y_continuous(name = "VEF1/CVF LIN", limits = c(0.7, 0.95)) +
  theme_bw()
s6d <- lln_panel(comp_m1, "fef2575_p5", "fef2575lln", "FEF25-75 (L/s) LIN")
figure_s6 <- ggarrange(s6a, s6b, s6c, s6d, labels = c("A", "B", "C", "D"))
figure_s6
Figure 11: Limite inferior da normalidade: nova equação (vermelho) vs. GLI2012 (azul) — Masculino Negro/Pardo.
Code 
figure_s7 <- ggarrange(s3c, s4c, s5c, s6c, labels = c("A", "B", "C", "D"))
figure_s7
Figure 12: VEF1/CVF LIN — comparação entre grupos (S3c, S4c, S5c, S6c).
Code 
figure_s8 <- ggarrange(s3a, s4a, s5a, s6a, labels = c("A", "B", "C", "D"))
figure_s8
Figure 13: VEF1 LIN — comparação entre grupos (S3a, S4a, S5a, S6a).

Comparação do LLN com a Amostra 2020

Code 
# FEV1
p_fev1 <- ggplot() +
  geom_smooth(data = comp, aes(height, fev1lln),  color = "blue") +
  geom_smooth(data = comp, aes(height, vef1_p5),  color = "red") +
  labs(x = "Estatura (cm)", y = "LIN VEF1")
p_fev1 + facet_wrap(~ sex + ethnic.f) + theme_classic(base_size = 16)

# FVC
p_fvc <- ggplot() +
  geom_smooth(data = comp, aes(height, fvclln),  color = "blue") +
  geom_smooth(data = comp, aes(height, cvf_p5),  color = "red") +
  labs(x = "Estatura (cm)", y = "LIN CVF")
p_fvc + facet_wrap(~ sex + ethnic.f) + theme_classic(base_size = 16)

# FVC — P50 e P5
p_fvc2 <- ggplot() +
  geom_smooth(data = comp, aes(height, fvcpred),  color = "blue") +
  geom_smooth(data = comp, aes(height, cvf_p50),  color = "red") +
  geom_smooth(data = comp, aes(height, fvclln),   color = "blue") +
  geom_smooth(data = comp, aes(height, cvf_p5),   color = "red") +
  labs(x = "Estatura (cm)", y = "CVF")
p_fvc2 + facet_wrap(~ sex + ethnic.f) + theme_classic(base_size = 16)

# FEV1/FVC
p_ratio <- ggplot() +
  geom_smooth(data = comp, aes(height, fev1fvclln),  color = "blue") +
  geom_smooth(data = comp, aes(height, vef1cvf_p5),  color = "red") +
  geom_smooth(data = comp, aes(height, fev1fvcpred), color = "blue") +
  geom_smooth(data = comp, aes(height, vef1cvf_p50), color = "red") +
  labs(x = "Estatura (cm)", y = "LIN VEF1/CVF")
p_ratio + facet_wrap(~ sex + ethnic.f) + theme_classic(base_size = 16)

# FEF25-75
p_fef <- ggplot() +
  geom_smooth(data = comp, aes(height, fef2575pred),  color = "blue") +
  geom_smooth(data = comp, aes(height, fef2575_p50),  color = "red") +
  geom_smooth(data = comp, aes(height, fef2575lln),   color = "blue") +
  geom_smooth(data = comp, aes(height, fef2575_lin),  color = "red") +
  labs(x = "Estatura (cm)", y = "FEF25-75 LIN")
p_fef + facet_wrap(~ sex + ethnic.f) + theme_classic(base_size = 16)
Figure 14: Comparação dos LLNs: GLI2012 (azul) vs. nova equação (vermelho), estratificado por sexo e etnia.
Figure 15: Comparação dos LLNs: GLI2012 (azul) vs. nova equação (vermelho), estratificado por sexo e etnia.
Figure 16: Comparação dos LLNs: GLI2012 (azul) vs. nova equação (vermelho), estratificado por sexo e etnia.
Figure 17: Comparação dos LLNs: GLI2012 (azul) vs. nova equação (vermelho), estratificado por sexo e etnia.
Figure 18: Comparação dos LLNs: GLI2012 (azul) vs. nova equação (vermelho), estratificado por sexo e etnia.

Base de Dados Expandida — x2005 + Mallozi

Code 
mal <- read_excel("%-Dados Mallozi/Mallozi EXTRAIDOS - Maio2021.xlsx")
mj  <- read_excel("%-Dados Mallozi/MHJ_maiores13_rio_branco.xlsx")

mal <- mal |>
  mutate(
    cidade  = "São Paulo",
    center  = 19,
    sex     = case_when(sexf == "Masculino" | sexf == "Male"   ~ 1L,
                        sexf == "Feminino"  | sexf == "Female" ~ 2L),
    ethnic  = 1,
    ethnic01 = 0,
    ethnicf  = "Branca"
  )

mj <- mj |>
  mutate(
    cidade = "Porto Alegre",
    center = 13
  )

malmj <- full_join(mal, mj) |>
  mutate(
    protocol = as.factor(protocol),
    ethnic   = as.factor(ethnic),
    ethnic01 = as.factor(ethnic01)
  )

xm <- full_join(x2005, malmj) |>
  mutate(
    sexf = case_when(
      sexf == "Masculino" ~ "Male",
      sexf == "Feminino"  ~ "Female",
      TRUE                ~ as.character(sexf)
    ) |> as.factor(),
    sex   = if_else(sexf == "Male", 1L, 2L),
    cidade = as.factor(cidade)
  )

write_xlsx(xm, "xm.xlsx")
Code 
ggplot(xm, aes(height, fvc,     col = protocol)) + geom_point(alpha = 0.1) + geom_smooth() + labs(title = "CVF")
ggplot(xm, aes(height, fev1,    col = protocol)) + geom_point(alpha = 0.1) + geom_smooth() + labs(title = "VEF1")
ggplot(xm, aes(height, fef2575, col = protocol)) + geom_point(alpha = 0.1) + geom_smooth() + labs(title = "FEF25-75")
ggplot(xm, aes(height, fev1fvc, col = protocol)) + geom_point(alpha = 0.1) + geom_smooth() + labs(title = "VEF1/CVF")
Figure 19: VEF1, CVF, FEF25-75 e VEF1/CVF em função da estatura na base expandida, por protocolo.
Figure 20: VEF1, CVF, FEF25-75 e VEF1/CVF em função da estatura na base expandida, por protocolo.
Figure 21: VEF1, CVF, FEF25-75 e VEF1/CVF em função da estatura na base expandida, por protocolo.
Figure 22: VEF1, CVF, FEF25-75 e VEF1/CVF em função da estatura na base expandida, por protocolo.