problem with aucint.inf.obs and missing data

[kirstenbergmann]

I have a dataset with missing concentration values (NA values) and get an error when trying to get AUCint (using aucint.inf.obs). The example runs fine if I replace the NA value with a number. The example also runs fine if I only request AUClast. I did try to include the option conc.na="drop" but this doesnt seem to solve the problem.

library('PKNCA')
conc <- data.frame(time=c(0,0.2833,0.5,1,2,3,4,5,6.0167,12,23.917),
            conc=c(1378.39,NA,1671.20,1553.99,1650.02,2044.83,1883.92,2045.73,1528.76,1701.77,1219.65),
                   SUBJID=rep(1,11))
dose <- data.frame(time=0,dose=5000,SUBJID=1)
my.conc <- PKNCAconc(conc,conc~time|SUBJID)
my.dose <- PKNCAdose(dose,dose~time|SUBJID)

my.interval <- data.frame(start=0,
                          end=24,
                          auclast=TRUE,
                          aucint.inf.obs=TRUE)


my.data <- PKNCAdata(my.conc, my.dose,intervals=my.interval)
my.results <- pk.nca(my.data)
#> Warning in pk.calc.aucint(conc = conc, time = time, start = start, end =
#> end, : Some interpolated/extrapolated concentration values are missing Time
#> points with missing data are: 0.2833

[billdenney]

Hi @kirstenbergmann,

It's just a typo in the interval column name for aucinf.obs. Replace your my.interval with the code below, and it will work.

my.interval <- data.frame(start=0,
                          end=24,
                          auclast=TRUE,
                          aucinf.obs=TRUE)

[kirstenbergmann]

Hi Bill

I dont want to calculate AUCinf but AUCtau, which in this case is not the same as AUClast as the last measurement is not at 24hours. I think aucint.inf.obs is what I want to use to get AUCtau. Below is an example without NA in the dataset:

library('PKNCA')
conc <- data.frame(time=c(0,0.2833,0.5,1,2,3,4,5,6.0167,12,23.917),
                   conc=c(1378.39,1501.23,1671.20,1553.99,1650.02,2044.83,1883.92,2045.73,1528.76,1701.77,1219.65),
                   SUBJID=rep(1,11))
dose <- data.frame(time=0,dose=5000,SUBJID=1)
my.conc <- PKNCAconc(conc,conc~time|SUBJID)
my.dose <- PKNCAdose(dose,dose~time|SUBJID)

my.interval <- data.frame(start=0,
                          end=24,
                          auclast=TRUE,
                          aucint.inf.obs=TRUE,
                          aucinf.obs=TRUE)

my.data <- PKNCAdata(my.conc, my.dose,intervals=my.interval)
my.results <- pk.nca(my.data)
as.data.frame(my.results)
#>    start end SUBJID            PPTESTCD      PPORRES exclude
#> 1      0  24      1             auclast 3.765216e+04    <NA>
#> 2      0  24      1                tmax 5.000000e+00    <NA>
#> 3      0  24      1               tlast 2.391700e+01    <NA>
#> 4      0  24      1           clast.obs 1.219650e+03    <NA>
#> 5      0  24      1            lambda.z 1.479558e-02    <NA>
#> 6      0  24      1           r.squared 6.287049e-01    <NA>
#> 7      0  24      1       adj.r.squared 2.574098e-01    <NA>
#> 8      0  24      1 lambda.z.time.first 6.016700e+00    <NA>
#> 9      0  24      1   lambda.z.n.points 3.000000e+00    <NA>
#> 10     0  24      1          clast.pred 1.268509e+03    <NA>
#> 11     0  24      1           half.life 4.684826e+01    <NA>
#> 12     0  24      1          span.ratio 3.820910e-01    <NA>
#> 13     0  24      1          aucinf.obs 1.200856e+05    <NA>
#> 14     0  24      1      aucint.inf.obs 3.775333e+04    <NA>

[billdenney]

@kirstenbergmann, sorry for misunderstanding the question. That is definitely a bug then!

[billdenney]

I just fixed this.