No output when trying to plot spline curves

[Sigrid Vikjord]

Dear Statalist,

I want to plot a spline curve of free fat mass (FFM) towards HRs from a Cox regression. I use the dataset and the code below. I have used a similar code before with other variables than FFM, and it has worked as it should. However, this time I am getting a strange output (copy-pasted below), with only missing values in the expression, and a resulting graph without anything else than a coordinate system. I have tried a workaround using

replace Ffm_NT4BLM=round(Ffm_NT4BLM,.5)

without any luck. Could anyone help me?

* Example generated by -dataex-. For more info, type help dataex
clear
input float(id enddate_bmi) double RegisStat long PartDat_NT4LuM double Ffm_NT4BLM
  1 22006 1     .    .
  2 22006 1     .    .
  3 22006 1     .   45
  4 22006 1     .    .
  5 22006 1     . 76.5
  6 22006 1 21214   66
  7 22006 1 21145 41.5
  8 17667 5     .    .
  9 22006 1     .   48
 10 16694 3     .    .
 11 21350 5     .    .
 12 22006 1     . 45.5
 13 22203 5     . 59.5
 14 22006 1     .    .
 15 16237 5     .    .
 16 22006 1     .    .
 17 16847 5     .    .
 18 22006 1     .    .
 19 22006 1     .   73
 20 22006 1     .    .
 21 22006 1     . 62.5
 22 22006 1     . 61.5
 23 22006 1     .    .
 24 22006 1     .    .
 25 15628 5     .    .
 26 22006 1     .   44
 27 13984 5     .    .
 28 22006 1     .    .
 29 20499 5     .    .
 30 22006 1 21095 48.5
 31 22006 1     . 58.5
 32 22006 1     .    .
 33 22006 1     . 37.5
 34 22006 1     .    .
 35 22006 1     .    .
 36 22006 1     .    .
 37 22006 1     . 48.5
 38 13710 5     .    .
 39 22006 1     .    .
 40 22006 1     .   35
 41 16420 5     .    .
 42 22006 1     .   66
 43 22006 1 21353 50.5
 44 22006 1 21087   67
 45 22006 1     .   55
 46 22006 1     .    .
 47 22006 1     .   65
 48 22006 1     .    .
 49 22006 1     .    .
 50 22006 1     . 44.5
 51 22006 1     . 42.5
 52 15294 5     .    .
 53 22006 1     .    .
 54 22006 1     .    .
 55 22006 1 21432 37.5
 56 22006 1     .   68
 57 22006 1     .    .
 58 22006 1     . 51.5
 59 22006 1 21440   45
 60 22006 1     . 43.5
 61 21776 5     .    .
 62 16724 5     .    .
 63 22006 1     . 39.5
 64 22006 1     . 64.5
 65 21046 5     .    .
 66 22006 1     .    .
 67 22006 1     .   45
 68 22006 1     .    .
 69 14715 5     .    .
 70 18246 3     .    .
 71 22006 1     .   37
 72 22006 1     .    .
 73 22006 1     . 41.5
 74 22006 1 21819   44
 75 22006 1     .    .
 76 22006 1     .    .
 77 17212 5     .    .
 78 22006 1     .    .
 79 22006 1     . 56.5
 80 22006 1 21111 79.5
 81 22006 1 21080 44.5
 82 22111 5     .   58
 83 22006 1 21800   71
 84 17577 5     .    .
 85 22006 1     .    .
 86 17943 5     .    .
 87 22006 1     .    .
 88 22006 1     .    .
 89 16875 5     .    .
 90 21929 5     .    .
 91 22006 1     .    .
 92 22203 5     .   45
 93 22006 1     .   51
 94 22006 1     .    .
 95 22006 1     .    .
 96 18550 5     .    .
 97 22006 1 21265   48
 98 22006 1     .    .
 99 22006 1     . 54.5
100 22006 1     . 37.5
end
format %td enddate_bmi
format %tdD_m_Y PartDat_NT4LuM
label values RegisStat RegisStat
label def RegisStat 1 "Bosatt", modify
label def RegisStat 3 "Utvandret", modify
label def RegisStat 5 "Død", modify

foreach num of numlist 3/7{
        preserve        
        stset enddate_bmi, id(PID_109925) failure(RegisStat==5) origin(time PartDat_NT4LuM) scale(365.25)
        mkspline spl_ffm = Ffm_NT4BLM, cubic nknots(`num') displayknots            
        stcox c.spl_ffm* i.Sex, vce(robust)
        estat ic
        levelsof Ffm_NT4BLM, local(levels)
        xblc spl_ffm*, covname(Ffm_NT4BLM) at(`levels') reference(54.8) eform gen(wbmi hr lb ub)
        keep if wbmi>26.0 & wbmi<118.0
        twoway rline lb ub wbmi, lp(solid) lc(red) lw(medium) || line hr wbmi, lc(blue) lp(solid) lw(medium) ///
        legend(off) ylabel(, nogrid format(%3.2f)) yticks() xlabel(26(1)118, format(%3.0f)) xticks() xtitle("{bf:BMI             (kg/m{sup:2})}") ytitle("{bf:HR (95% CI) for all-cause mortality}") ///
        yline(1, lc(black) lp(dash) lw(medthin)) name(no_copd_death_spl_ffm_knot`num', replace) saving("C:\.gph",             replace)
        restore
}

Parts of output:

Ffm_NT4BLM    exp(xb)    (95% CI)
26.5            .        (  .-  .)
27              .        (  .-  .)
27.5            .        (  .-  .)
28              .        (  .-  .)
28.5            .        (  .-  .)
29              .        (  .-  .)
29.5            .        (  .-  .)
30              .        (  .-  .)
30.5            .        (  .-  .)
31              .        (  .-  .)
31.5            .        (  .-  .)
32              .        (  .-  .)
32.5            .        (  .-  .)
33              .        (  .-  .)
33.5            .        (  .-  .)
34              .        (  .-  .)
34.5            .        (  .-  .)
35              .        (  .-  .)
35.5            .        (  .-  .)
36              .        (  .-  .)
36.5            .        (  .-  .)
37              .        (  .-  .)
37.5            .        (  .-  .)

[Sigrid Vikjord]

I fixed the error: I needed to round the mean of the dependent variable up to the nearest integer.