What is your approach to dosing and duration of rituximab for refractory lupus nephritis not responsive to cyclophosphamide or mycophenolate?

First

The question specifically states "not responsive." It would be highly improbable not to have some improvement with these therapies in conjunction with using hydroxychloroquine (HCQ), steroids, ACE inhibitors, ARBs, and sodium-glucose cotransporter 2 (SGLT2) inhibitors in an adherent patient.

What type of lupus nephritis? What did the biopsy actually show?

My first question would be, "Was the renal biopsy read properly?" I'd ensure a nephropathologist read the original biopsy (I like using Johns Hopkins pathology).

Also, all lupus nephritis is not the same. Was this class III, class IV, class V, membraneproliferative? Were there signs of a poor prognosis, such as cellular crescents and fibrinoid necrosis? If so, how severely were the glomeruli affected? I'd want to assess if there were complicating issues like tubulointerstitial disease, fibrosis, and arteriosclerosis.

Adherence to therapy?

Adherence to therapy must be assessed. I would absolutely have been checking MMF and through whole-blood HCQ drug levels to ensure adherence. 

Rebiopsy is in order

If all the above was used and there was "no response," then a biopsy is needed to reassess the lesion (Morales et al., PMID 33081588). Rapid fibrosis, arteriosclerosis, FSGN, and other lesions can occur in this scenario of "no response." I would ask an excellent nephropathologist to read the specimen. 

My method: assuming there were some improvements with treatment, excellent adherence, and ongoing treatable LN without any untreatable lesions

Let's assume there was some improvement (which would be much more common) adherence was not a problem, and appropriate BP treatments and proteinuria therapies were being used:

Use FDA-approved LN therapies

I would not use a CD-20-targeted B-cell depleting biologic next. I prefer to try FDA-approved drugs for lupus nephritis (LN) rather than jumping to rituximab (RTX). 

We have two FDA-approved drugs for LN: belimumab (BEL, Benlysta) and voclosporin (VOC, Lupkynis).

Of course, I would not want to retry cyclophosphamide, but I would use mycophenolate mofetil (MMF) with these agents.

Using mycophenolic acid in PPI-dependent patients

If the patient were on a proton pump inhibitor, I would use mycophenolic acid (MPA) instead. PPIs can significantly reduce the bioavailability of MMF. I prefer MPA in PPI-dependent patients (like our patients with esophageal dysmotility). 

Using combination therapy for an optimal outcome

By this time, we have wasted a lot of time. "Time is kidney," is the saying. We must assume that this patient has already permanently lost many additional glomeruli by now. We cannot take chances on losing more and need to give this patient therapy that has the best chance of achieving remission off steroids.

Therefore, I would prefer using the combination of BEL plus VOC (Baum et al., PMID 37303434) along with MMF/MPA + HCQ + proteinuria lowering agents. I have had great success with this approach. BEL addresses the B-cells and extra-renal manifestations (and BEL is a disease-modifying agent in LN and SLE). Then, VOC is usually quite impressive in reducing proteinuria rapidly via its effects on the podocytes. 

This combination would give this patient the best chances for achieving a response (and hopefully renal remission), and going down on (and hopefully off) steroids. 

Insurance coverage of BEL and VOC

NOTE: I have not had trouble getting BEL and VOC in a patient. We are using it within the FDA-approved package inserts. Both agents must be used with MMF along with other standard lupus therapies. BEL and VOC are both standard therapies for SLE and LN. 

Word of caution with steroids

I would try to taper them as much as possible. There is little evidence that they do much for LN after initial treatment with IV methylprednisolone and high-dose steroids with a taper. In other words, I would not be increasing steroids for possible flares of LN.

Using obinutuzumab instead of rituximab

Now, in that very rare instance where we are truly dealing with a resistant biopsy-proven LN that has failed all of the above and the patient is a treatment resistant to those therapies discussed above, I would prefer using obinutuzumab (OBI, Gazyva) over rituximab. The high response rate of using OBI in LN, along with the impressive side effect profile when compared to placebo + standard therapy, is impressive (Furie et al., PMID 34615636). I would use 1000 mg given 2 weeks apart and monitor closely at least monthly. I would most likely use a calcineurin inhibitor (again, attacking the lesion with several complementary modes of action). I would repeat this 2-week treatment six months later.

And if rituximab must absolutely be used

If you are unable to get OBI, then I would use RTX 1000 mg 2 weeks apart (then repeat this regimen 4-6 months later, depending on the response). I would use a calcineurin inhibitor as well. Other studies have used a dose of 375 mg/m2 weekly for four weeks (Melander et al., PMID 19261822). To my knowledge, I do not think there are any head-to-head comparisons of these two regimens to prove the superiority of one over the other.