How can show all legend labels on a bar plot, without dropping them when data does not exist?

Question

I am using patient dataset to determine which medications (amiodarone & digoxin) they are taking.

Here is the classifer I am using

topcat$aARGroup <- with(topcat, ifelse((is.na(amiodarone) & is.na(digoxin)), 'Missing data',
                                 ifelse((amiodarone == 1 & digoxin == 0), 'Amiodarone',
                                   ifelse((amiodarone == 0 & digoxin == 1), 'Digoxin',
                                     felse((amiodarone == 1 | digoxin == 1), 'Combination', 'No usage')))))
topcat$aARGroup <- factor(topcat$aARGroup, levels = c('Amiodarone','Digoxin','Combination','None','Missing data'))

Unfortunated all data are 'NA' since it is not recorded. So it is 100% "Missing data". I am trying to generate a bar plot and have all medication claases displayed in the legend, eventhough all classes except "missing data" are empty.

ggplot(topcat, aes(x = aARGroup))+
  theme(axis.text.x=element_blank(), axis.ticks.x=element_blank())+
  geom_bar(aes(y = ..prop.., fill = factor(..x..)), stat="count") +
  geom_text(aes( label = scales::percent(..prop.., accuracy = 0.1),
                 y= ..prop.. ), stat= "count", vjust = -.5, size = 5) +
  scale_y_continuous(labels=scales::percent) +
  labs(y="Percentage",x="Anti Arrhythmic drugs")+
  scale_x_discrete(limits=c('Amiodarone','Digoxin','Combination','No usage','Missing data'))+
  scale_fill_discrete(labels=c('Amiodarone','Digoxin','Combination','No usage','Missing data'))+
  theme(text=element_text(size=18), legend.position = "bottom", legend.title = element_blank())+
  theme(strip.text=element_text(size=11))

So the code does not produce the expected results.

1. It does not show all classes
2. It shows all patients are taking Amiodarone, when underlying data all "Missing data"

Here is the output

enter image description here

Here is the dataframe view

enter image description here

Thank you for your help!