Combination therapies have always seemed to work best for me in patients who are tough to treat.

Omalizumab or cyclosporine may be good options. While you will definitely need to check bloodwork routinely while your patients are on cyclosporine, you can obtain control quickly, and after your patient has been controlled for 4-6 months, you can wean the dose to off and maintain benefit. It might take that long simply to get omalizumab approved for this indication!
The old mainstay used to be hydroxyzine. I think hydroxyzine and cetirizine (most active metabolite of hydroxyzine) are still the most effective H2 blockers for urticaria, but often the doses used are too low. FDA-approved dose of hydroxyzine is 400 mg/day (100q6). I try to push the dose up without too much sedation, and get them to 50-100 mg at least at night, and some during the day as tolerated, and tell them not to drive if necessary. The sedative effects tend to decrease with chronic use of these antihistamines. Since serum levels of cetirizine are equivalent for cetirizine 10 mg qd and hydroxyzine 50 mg qd, I often prescribe cetirizine 10-20mg during the day. (If sedated on cetirizine during the day, I use loratadine, fexofenadine etc.) If poor response, add doxepin, again with caution about sedation. Can try H1 blocker, but helps in very few patients. If still no success, try leukotriene inhibitor. Zileuton probably works best. Then try cyclosporin, omalizumab, hydroxychlorquine, and dapsone.
We’ve had some good success with Plaquenil 400 mg qd, which is the immunomodulator with the least side effects. One recalcitrant 70 y/o lady failed 2 mo. trial of that, has HTN so I used Imuran 50 mg qd and she is 100% symptom free (monitoring CBC and LFT’s q 6 wk). One drug that I have seen work in many patients who have failed others is Doxepin. As an antihistamine, it is about 800 times as potent as Benadryl. Though it is not easy to take because of the side-effects, if you start at a very low dose and gradually work your way up (starting QHS) the results can be very impressive. I have successfully gotten many patients off of steroids this way. I always give the patients strong warnings about the side-effects of the drug (extreme sedation, dry mouth, dry eyes, constipation etc.), but they generally seem more willing to try it than cyclosporine, Imuran and others.
Personally, I remain unconvinced that anything other than oral steroids and H1 blockers work. In this case I would probably try Xyzal 5 mg BID, hydroxizine HCl or Benadryl prn for breakthrough sx, and change the prednisone to methylpred. in case the pt. is a slow acetylator (? I think that’s right). I would back up the oral steroid to 24 mg of medrol daily (=30 of PDN) until under control and then taper the alternate day by 4 mg every other day: 24/20/24/16/ etc. until the patient is on 24 qod. At that point I would taper slowly, every week or so while continuing maximal antihistamines.
I’ve also had some success with Plaquenil. Interestingly enough I have had success with anxiolytics in some patients who swear their sxs are stress induced. Good letter to the editor in this months JACI regarding Alprazolam mechanism and use in severe refractory CU. First, recognize that no agent will be 100% successful, which definitely applies to any and all antihistamines. Once urticaria has been classified as refractory to antihistamines (and I would include montelukast here, since it is safe and familiar), then move onto more exotic therapies. Dapsone can work well in some and has the possibility for drug-free remission. Colchicine, hydroxychloroquine, and sulphasalazine have proven less reliable for me. More reliable but more expensive or monitoring-intensive meds include cyclosporine, tacrolimus, and mycophenolate. I like mycophenolate because it’s slightly safer, has 2 different formulations available, and can work well in some intractable cases. I suspect it’s slightly less effective than CyA or tacrolimus, though. It probably works in a similar fashion to azathioprine but with fewer side effects.
Failing those, you can try methotrexate, IG, phototherapy (requiring referral to a willing dermatologist), androgens, cyclophosphamide (including oral regimens), gold, or even plasmapheresis. There are reports on nifedipine, anticoagulants, oral beta agonists, and theophylline, but I am skeptical if these agents are worth trying. Newer (read: expensive) agents you could try include omalizumab and etanercept (and related agents).
Finally, you must think about thyroid autoimmunity. A patient with evidence of autoimmune thyroiditis and who is euthyroid by TSH might occasionally experience remission of the urticaria with exogenous, suppressive doses of thyroxine. Most things on this list, when carefully selected and executed are safer for the patient and more elegant than systemic steroids.