One of the comments to the last post mentioned that a way to provide the benefits of a combination therapy is to make a drug that hits more than one target. For a while there, this idea was a bit out of fashion, but I think it's been making a comeback.
There are a number of multiple-target drugs on the market already, and the list even includes some where this profile was arrived at deliberately. (That was a joke. I think.) Examples of the planned category include Tracleer (bosentan), a dual endothelin receptor antagonist for pulmonary hypertension, rupatadine, a dual histamine/PAF antagonist for allergy (not yet approved in the US, I think), and Cymbalta (duloxetine), an antidepressant which affects both serotonin and noradrenaline reuptake.
Examples of the non-deliberate class include all sorts of drugs marketed before the 1980s or so. (That's about the time that target-based drug design really took over, as opposed to "see what it does in vivo".) In some therapeutic classes, this is the norm - the activities of CNS drugs in general are known to be extraordinarily messy. The antipsychotics, for example, hit so many receptors that it's basically impossible to figure out just how they work. A more recent example is the cardiovascular drug Pletal (cilostazol), which is both a PDE-III inhibitor and an adenosine uptake inhibitor. Other PDE-III inhibitors didn't work nearly as well, so there was clearly something else pitching in. (There may well be something similar at work with Lipitor, to pick a drug that everyone's actually heard of, although no one's quite sure what the extra activity is.)
There are all sorts of other dual-acting drugs being looked at in earlier phases of development. For the most part, they're going after similar receptor subtypes or related enzymes, since that's where you're most likely to get the cross-reactivity. (A good example would be the PPAR alpha-gamma ligands that many companies have been trying to develop for diabetes.) But the biggest area of multiple-action drugs now is cancer.
You might not know it from reading the popular press, though. A lot of reporters are still a generation behind, going on about the new breed of incredibly selective targeted cancer drugs. Problem is, it's turning out that some of those incredibly selective drugs work only on incredibly small numbers of cancer patients, which is not what everyone had in mind. Over the last few years, efforts have shifted to making drugs that hit a slew of potential cancer targets simultaneously, in hopes that this will show more efficacy, and these are just coming to the FDA now.
Many, many cancer targets are from a large family of broadly similar enzymes (the kinases), so getting multiple activities isn't really all that hard. In fact, getting selective kinase inhibitors was the hard part - looking back, had we but known, we all could have probably skipped that step and gone right to the blunderbusses. But the fear was that these compounds would be too toxic (yeah, even for cancer therapy), so selectivity got priority. Now that it turns out that we don't need to be so picky, it's also becoming clear that the multiple-kinase drugs are tolerated a lot better than we thought. You'll see the word "targeted" thrown around when these agents are discussed, but it should have quotes around it.