In R in prefiltering in deseq2 analysis which should i use: rowSums(counts(dds))>=patients or patients/2?

Question

I'm using Deseq2 analysis and I have 52 samples and on the stage of prefiltering I'm not sure if I have to use:

keep <- rowSums(counts(dds)) >= patients_number
dds2 <- dds[keep,]
dds2

or

keep <- rowSums(counts(dds)) >= patients_number/2
dds2 <- dds[keep,]
dds2

My original dimensions of my samples were:

class: DESeqDataSet 
dim: 62758 52 
metadata(1): version
assays(1): counts
rownames(62758): ENSG00000000005 ENSG00000000419 ... __not_aligned __alignment_not_unique
rowData names(0):
colnames(52): V3 V32 ... V80PLUS9 VRM6
colData names(2): Groups Newgroup

And after trying both patients_number or patients_number/2 the results were respectively:

dim: 19202 52 and dim: 21984 52 

But I don't know which to keep.