dat.anker2025: Cardiology Example Data with Subgroup and Interaction Effects

In metadat: Meta-Analysis Datasets

Cardiology Example Data with Subgroup and Interaction Effects

Description

Treatment effects, overall and in male/female patient subgroups, as well as treatment-by-subgroup interaction effects in six randomized, placebo-controlled trials.

Usage

dat.anker2025

Format

The data frame contains the following columns:

study	character	study identifier
year	integer	study year
intv.n	integer	number of patients in the intervention group
placebo.n	integer	number of patients in the placebo group
iv.iron	factor	the intravenous (IV) iron formulation investigated
size	factor	study size (“small” or “large”, based on total number of patients)
followup	numeric	followup duration (months)
baseline.age	numeric	mean age at baseline (years)
baseline.lvef	numeric	mean left ventricular ejection fraction (LVEF) at baseline (%)
baseline.hgb	numeric	mean haemoglobin at baseline (g/dl)
baseline.ft	numeric	mean ferritin at baseline (\mug/l)
baseline.tsat	numeric	mean transferritin saturation (TSAT) at baseline (%)
total.n	integer	total number of patients
total.logrr	numeric	overall effect (log-RR)
total.se	numeric	standard error of overall effect
female.n	integer	number of female patients
female.logrr	numeric	effect (log-RR) in females
female.se	numeric	standard error of effect in females
male.n	integer	number of male patients
male.logrr	numeric	effect (log-RR) in males
male.se	numeric	standard error of effect in males
sex.logrrr	numeric	treatment-by-subgroup interaction (log-RRR)
sex.se	numeric	standard error of treatment-by-subgroup interaction

Details

Anker et al. (2025) analyzed the effects of intravenous (IV) iron therapy observed in six randomized, placebo-controlled trials. The primary endpoint was a composite of (recurrent) heart failure (HF) hospitalizations and cardiovascular (CV) death, and treatment effects were quantified in terms of risk ratios (RRs). Individual participant data were available for five trials, and analyses were harmonized to match the analysis of the sixth trial (the Ironman study). Meta-analyses were then performed based on logarithmic RRs (log-RRs).

Besides investigation of the overall effect, it was of interest to what extent male and female patients benefited, and whether effects differed between both patient subgroups. The data set includes effect estimates within each study's male and female subgroups (in terms of log-RRs), as well as estimates of the difference between the subgroups (in terms of logarithmic ratios of RRs (log-RRRs). Technically, the RRRs then constitute treatment-by-subgroup interaction effects.

Concepts

medicine, cardiology, incidence rates

Author(s)

Christian Roever, [e-mail]

Source

Anker, S. D., Karakas, M., Mentz, R. J., Ponikowski, P., Butler, J., Khan, M. S., Talha, K. M., Kalra, P. R., Hernandez, A. F., Mulder, H., Rockhold, F. W., Placzek, M., Röver, C., Cleland, J. G. F., & Friede, T. (2025). Systematic review and meta-analysis of intravenous iron therapy for patients with heart failure and iron deficiency. Nature Medicine, 31(8), 2640-2646. ⁠https://doi.org/10.1038/s41591-025-03671-1⁠

Examples

# show data
dat.anker2025
## Not run: 
library(bayesmeta)

# specify heterogeneity (tau) prior density (half-normal(0.5))
HN05 <- function(t){dhalfnormal(t, scale=0.5)}

#######################################
#  reproduce primary analysis (Fig.2)
es.primary <- escalc(measure="IRR", yi=total.logrr, sei=total.se,
                     ni=total.n, slab=study, data=dat.anker2025)
bma01 <- bayesmeta(es.primary, tau.prior=HN05)
forestplot(bma01, expo=TRUE, xlog=TRUE,
           xlab="rate ratio (RR)", title="HF hospitalizations / CV mortality")

######################################
#  reproduce sex interaction analysis
#  (Tab.2, Supplementary Fig.4)
es.sex <- escalc(measure="GEN", yi=sex.logrrr, sei=sex.se,
                 ni=total.n, slab=study, data=dat.anker2025)
bma02 <- bayesmeta(es.sex, tau.prior=HN05)
forestplot(bma02, expo=TRUE, xlog=TRUE,
           xlab="ratio of RRs (RRR)", title="sex interaction (females/males)")

###########################################
#  reproduce male/female subgroup analyses
#  (Tab.2, Supplementary Fig.12)
es.female <- escalc(measure="IRR", yi=female.logrr, sei=female.se,
                    ni=female.n, slab=study, data=dat.anker2025)
es.male   <- escalc(measure="IRR", yi=male.logrr, sei=male.se,
                    ni=male.n, slab=study, data=dat.anker2025)
bma03a <- bayesmeta(es.female, tau.prior=HN05)
bma03b <- bayesmeta(es.male, tau.prior=HN05)
forestplot(bma03a, expo=TRUE, xlog=TRUE,
           xlab="rate ratio (RR)", title="female subgroup")
forestplot(bma03b, expo=TRUE, xlog=TRUE,
           xlab="rate ratio (RR)", title="male subgroup")

## End(Not run)

metadat documentation built on April 29, 2026, 5:10 p.m.