1. Using forestplot() function

1.1. Plot 1

Prepare data

# Create tibble
forest_data <- tibble(
  exposure = c(
    "Full cohort",
    "  1st trimester",
    "  2nd trimester",
    "  Early 3rd trimester",
    "  Late 3rd trimester",
    "Subpopulation analyses",
    "  1st trimester",
    "  2nd trimester",
    "  Early 3rd trimester",
    "  Late 3rd trimester"
  ),
  mean  = c(NA, 1.20, 1.31, 1.67, 1.32, NA, 1.11, 1.28, 2.06, 1.48), # point estimates (ORs)
  lower = c(NA, 0.95, 1.06, 1.17, 1.04, NA, 0.83, 0.99, 1.29, 1.10),
  upper = c(NA, 1.51, 1.63, 2.38, 1.68, NA, 1.48, 1.66, 3.31, 1.99),
  aOR_label = c(NA, "1.20 (0.95–1.51)", "1.31 (1.06–1.63)", "1.67 (1.17–2.38)",
                "1.32 (1.04–1.68)", NA, "1.11 (0.83–1.48)", "1.28 (0.99–1.66)",
                "2.06 (1.29–3.31)", "1.48 (1.10–1.99)"),
  is_summary = c(TRUE, FALSE, FALSE, FALSE, FALSE, TRUE, FALSE, FALSE, FALSE, FALSE) # TRUE = summary row (bold), FALSE = normal row
)
# Create table text with header row
tabletext <- rbind( #to stack rows together
  c("Exposure status", " ", "aOR (95% CI)"),  # first row with column titles
  cbind(
    forest_data$exposure, " ", forest_data$aOR_label # binds 3 columns
    )
)

So, the forest plot will have 3 columns, the middle column is for plot

# Horizontal lines
h_lines <- list(
  "2" # # draw a line after header row
  = gpar(lty = 2, # dashed line
              col = "grey"))                           

# Forest plot 
forestplot(
  labeltext = tabletext, 
  mean  = c(NA, forest_data$mean), # point estimates, add NA for header row
  lower = c(NA, forest_data$lower), # lower CI, NA for header
  upper = c(NA, forest_data$upper),
  is.summary = c(TRUE, forest_data$is_summary), # mark header and subgroup titles bold
  zero = 1, # null value = 1
  xlog = TRUE, # plot with log(OR) values
  xlab = "aOR (95% CI)", # x axis label
  boxsize = 0.3,
  lineheight = unit(0.6, "cm"), # verticle spacing between rows
  xticks = c(0.5, 1, 1.5, 2.0, 3.0), # x axis tick marks
  clip = c(0.5, 3.0), # clipping area
  graph.pos = 2, # specify plot in 2nd column
  graphwidth = unit(50, "mm"), #width of plot column
  colgap = unit(2, "mm"), # gap between table and plot
  col = fpColors(box = "#1f77b4", line = "#1f77b4", summary = "black"),
  txt_gp = fpTxtGp(
    label = gpar(fontsize = 12),
    ticks = gpar(fontsize = 13),
    xlab = gpar(fontsize = 14)
  ),
  align = c("l", "c", "c"), # align columns
  hrzl_lines = h_lines
)

1.2. Plot 2

forest_data_1 <- tribble(
  ~exposure, ~mean, ~lower, ~upper, ~is_summary,
  "Close-to-delivery exposures", NA, NA, NA, TRUE,   # header
  "Any antibiotics",             NA, NA, NA, TRUE,   # subgroup
  "No adjustment for GBS risk factors", 1.70, 1.31, 2.19, FALSE,
  "Adjusted for GBS risk factors",      1.29, 0.99, 1.67, FALSE,
  "GBS treatment",              NA, NA, NA, TRUE,   # subgroup
  "No adjustment for GBS risk factors", 1.76, 1.34, 2.30, FALSE,
  "Adjusted for GBS risk factors",      1.34, 1.02, 1.76, FALSE,
  "Penicillins",                NA, NA, NA, TRUE,   # subgroup
  "No adjustment for GBS risk factors", 1.54, 1.12, 2.13, FALSE,
  "Adjusted for GBS risk factors",      1.17, 0.84, 1.62, FALSE
)


# ----------------------------
# Table text for forestplot
# ----------------------------
tabletext_1 <- cbind(
  c("Exposure", forest_data_1$exposure),
  c("", rep("", nrow(forest_data_1))),
  c("aOR (95% CI)",
    "", "",
    "1.70 (1.31 – 2.19)",
    "1.29 (0.99 – 1.67)",
    "",
    "1.76 (1.34 – 2.30)",
    "1.34 (1.02 – 1.76)",
    "",
    "1.54 (1.12 – 2.13)",
    "1.17 (0.84 – 1.62)"
  )
)

#pdf("forest_plot.pdf", width = 6, height = 4)  
# ----------------------------
# Forest plot
# ----------------------------
forestplot(
  labeltext = tabletext_1,
  mean  = c(NA, forest_data_1$mean),
  lower = c(NA, forest_data_1$lower),
  upper = c(NA, forest_data_1$upper),
  is.summary = c(TRUE, forest_data_1$is_summary),
  zero = 1,
  xlog = TRUE,
  xlab = "aOR (95% CI)",
  boxsize = 0.3,
  lineheight = unit(0.6, "cm"),
  xticks = c(0.5, 1, 1.5, 2.5),
  clip = c(0.5, 2.5),
  graph.pos = 2,
  graphwidth = unit(50, "mm"),
  colgap = unit(2, "mm"),
  col = fpColors(box = "#1f77b4", line = "#1f77b4", summary = "black"),
  txt_gp = fpTxtGp(
    label = gpar(fontsize = 11),
    ticks = gpar(fontsize = 12),
    xlab = gpar(fontsize = 13)
  ),
  align = c("l", "c", "c"),
  hrzl_lines = h_lines
)

#dev.off()

2. Using forest() function

Prepare data (Odds ratio scale)

dt <- tribble(
  ~`Antibiotic exposure status`,              ~est,  ~low,  ~hi,
  "Full cohort",           NA,    NA,    NA,
  "1st trimester",        1.20,  0.95,  1.51,
  "2nd trimester",        1.31,  1.06,  1.63,
  "Early 3rd trimester",  1.67,  1.17,  2.38,
  "Late 3rd trimester",   1.32,  1.04,  1.68,
  "Subpopulation analyses", NA,   NA,    NA,
  "1st trimester",        1.11,  0.83,  1.48,
  "2nd trimester",        1.28,  0.99,  1.66,
  "Early 3rd trimester",  2.06,  1.29,  3.31,
  "Late 3rd trimester",   1.48,  1.10,  1.99
)
# Indent (if they have estimates, i.e. are not main headers)
dt$`Antibiotic exposure status` <- ifelse(is.na(dt$est),
                      dt$`Antibiotic exposure status`,
                      paste0("   ", dt$`Antibiotic exposure status`))
# Add blank column for forestploter plotting area
dt$` ` <- paste(rep(" ", 20), collapse = " ")

# Add formatted aOR (95% CI) column
dt$`aOR (95% CI)` <- ifelse(
  is.na(dt$est), "",
  sprintf("%.2f (%.2f to %.2f)", dt$est, dt$low, dt$hi)
)

# Compute SE for
dt$se <- ifelse(is.na(dt$est), NA, (log(dt$hi) - log(dt$est)) / 1.96)

# Build forest plot
p_1 <- forest(dt[, c("Antibiotic exposure status", " ", "aOR (95% CI)")],
            est = dt$est,
            lower = dt$low,
            upper = dt$hi,
            sizes = 0.5,
            ci_column = 2,
            ref_line = 1,
            arrow_lab = c("Antibiotics protective", "Antibiotics harmful"),
            xlim = c(0.5, 3.0),
            ticks_at = c(0.5, 1, 1.5, 3)
)

plot(p_1)