I use GAPDH and/or beta-Actin. Both typically give me CT around 16 cycles. I cannot see a reason why either would not work well for whole blood, but just to be accurate in my response, I have only ever used these primers for T-cells and Macrophages isolated from whole blood.
You ask a great question that has multiple answers, and that will require you to make a judgment call.
The first answer is tradition. Daniel, as mentioned above, uses traditional housekeeping genes such as GAPDH and beta-Actin. Also utilized in Western Analysis, you can find these housekeeping genes in a multitude to studies. If you are most comfortable with using what is widely used, then this may be the route to go.
The second answer is "newer" housekeeping genes. Recent studies have identified that GAPDH is problematic as a housekeeping gene because its levels fluctuate within cells and conditions, and it varies across tissue types (see openwetware.org link below, which also has links to the primary literature). One paper has found that GAPDH failed completely as a housekeeping gene in providing quantification of gene expression in cells that were undergoing differentiation, while genes encoding ribosomal proteins worked well (see BMCGenomics link). A very nice meta-analysis of housekeeping genes across microarray data sets (see PLOS ONE link below, I think it's worth your time to read!) showed that genes encoding ribosomal protein subunits were the most stably expressed across different tissues, and this has really picked up steam. Thus, if you want to use the most recent evidence, you may consider a ribosomal protein subunit gene.
The third answer is to combine multiple reference genes, even potentially a ribosomal subunit gene and GAPDH or GAPDH and ACTB. By averaging multiple reference genes (see Pubmed link below, which you should also check out at least Figure 5), you may average out any deficiencies in a single one and strengthen your analysis.
At the end of the day, if you are doing the same tissue type (i.e. human whole blood) you will probably be OK with GAPDH or ACTB. However, this is a hot topic, and I at least want to introduce you to some of the interesting possibilities and literature.