But Doc I'm in Pain!

What would you do if you had chronic pain and a GFR of 90? Would you think twice about a two-month (or longer) course of NSAIDs? What if your GFR was 60? Or 30? As society grows increasingly fearful of chronic opioid therapy, what are we, PCPs, and patients to do? The literature is somewhat sparse, not surprisingly, on the broader risks of NSAIDs to our beloved beans. Especially in the post-phenacetin combination analgesic era.

Last year, the results of the FDA-mandated non-inferiority safety trial of celecoxib were published. Focusing on CV/GI outcomes, it is not a renal trial. This trial has also been reviewed (with a nice visual abstract) in NephJC. 

Some basic info about the trial:

•  ~24,000 pts with elevated CVD risk and OA (~90%) or RA (~10%) pain were randomized to celecoxib, naproxen, or ibuprofen. 
• Mean daily doses: celecoxib ~200 mg, naproxen ~850 mg, ibuprofen ~2000 mg. ~1/2 of patients were also on ASA for CV protection. 
• Esomeprazole was given for GI protection though adherence was not reported.
• Baseline mean Cr was 0.9 +/- 0.2 mg/dL. Patients with CR > 1.7 (M), 1.5 (F) were excluded. 
• Mean treatment duration 20 +/- 16 months with mean follow-up 34 +/- 13 months. 
• ~2/3 stopped drug at some point and ~1/4 were lost in follow-up. 

Results: Celecoxib was "non-inferior" in terms of CV risk compared to ibuprofen and naproxen. The primary composite outcome of CVD, MI, CVA happened to 2.3-2.7% of the ITT groups. In other words, celecoxib is not worse for your heart than ibuprofen or naproxen. There may be a reduced risk of serious GI events with celecoxib as well.

3° outcomes included adjudicated “clinically significant renal events," which were:

• Cr ≥2.0 mg/dL with increase of Cr by ≥ 0.7 mg/dL.
• Admission for AKI (doubling Cr, AKI + K > 6, or requiring RRT).

Renal Outcomes in ITT Analysis (Nissen et al. 2016 NEJM 375 (26):2519-2529).

On-Treatment Analysis -- HR for renal outcomes were: celecoxib vs. naproxen 0.66 (0.44-0.97) and celecoxib vs. ibuprofen 0.54 (0.37-0.80).

So celecoxib may be safer from a renal standpoint than ibuprofen or naproxen in patients without baseline CKD. Note the moderate dose of celecoxib used and high rates of drop out. In addition, significant GFR losses over a relatively short period of time (e.g. Cr 0.9 to 1.5) may have been missed.

What about patients with CKD?

Cohort studies in patients with CKD are mixed and limited by the usual concerns. A recently published prospective cohort study of ~4,000 patients with rheumatoid arthritis compared eGFR trajectories in those having taken NSAIDs (~2,700) to NSAID naïve patients (~1,300). NSAID use was not associated with more rapid CKD progression in those with eGFR > 30 (using Cockgroft-Gault) though a very small number of patients with eGFR < 30 did have more rapid progression on NSAIDs.

Importantly, other cohort studies have shown that higher dose NSAID use may be associated with faster GFR decline and that selective COX-2 inhibitors may not be any safer than non-selective NSAIDs. The latter notion may contradict the study discussed above.

NSAIDs can reduce GFR acutely, worsen hypertension, or provoke hyperkalemia. But this is a concern only for a minority of most CKD patients right? Can close monitoring of patients and striving to use the lowest dose reduce our concerns and allow greater use of these effective medications in patients with CKD? Especially in those patients with "lower-risk" CKD? Given our limited options otherwise, I hope so.

For those interested, a recent review in AJKD about pain control in CKD (and a notable response) provide additional information and are worthy reads.

Rob Rope
Stanford Nephrology Fellow

NSAIDs: new tricks from an old drug

A 62-year-old man with CKD stage 3 secondary to presumed hypertension presented with new-onset hypercalcemia. PTH level was low at 11 and abdominal imaging revealed a right renal mass that later was confirmed to be a renal cell carcinoma.

The most common cause of hypercalcemia of malignancy is secretion of PTH-related peptide, followed by increased bone resorption due to bone mets and increased 1-alpha hydroxilase activity by the tumor (similar to granulomatous disease).

His workup was remarkable for iCa 6.3, total Calcium 11.5, PO4 4.6, ACE level 36 and alk phos 109. Negative SPEP/UPEP, no elevation of 1,25 vitamin D and a mildly elevated PTHrP 49 pg/mL (14-27). One interesting aspect of PTHrP is that the most common assay measures the carboxy-terminal fragment, which is usually removed by the kidney and might be abnormality elevated in renal insufficiency. On our patient, his hypercalcemia (~11-13) was associated with worsening renal function and creatinine peaked at 4.3 mg/dl. Despite hydration and ibandronate 6mg IV, his hypercalcemia persisted at ~11-12 mg/dL (72 hours afterwards).

For further workup, another assay was performed with the amino-terminal fragment of PTHrP and revealed a low value (0.8). Lastly, there was no evidence of metastasis on bone scans. So what was the cause of the hypercalcemia in our patient and how can we get it under control?

Solid tumor cells are believed to secrete other bone-resorbing factors that mediate hypercalcemia through increased production of prostaglandins E (PGE). PGE is a potent stimulator of bone reabsorption in vitro and there is both animal and human data suggesting that PGE is an important contributor of hypercalcemia in some malignancies.

The most interesting aspect of it is that treatment involves administration of NSAIDS, which can block the production of PGE and resolves hypercalcemia. The original paper dates back to 1975 at the NEJM and there has been not many reports after that. Nonetheless, the team taking care of the patient started him on indomethacin 25mg q8hours as a last attempt to control his calcium before removal of the renal mass was performed. The results were astonishing as you can see in the graph above.

So despite all the negative effects of NSAIDs in the kidney, we have now a legitimate use in certain patients that have hypercalcemia 2/2 malignancy. Few questions still remain: how common is PGE-mediated hypercalcemia? Should we be checking for PGE levels as part of the workup? It seems that in patients that are not being considered for surgical resection and that did not respond to biphosphonates, a trial of NSAIDS could be warranted. Especially since the response is so quick!

Please share your experience as well since this was truly the first time I heard of this correlation. Kudos to the team taking care of this patient (led by Dr Mount) and Omar's grand rounds' presentation.
(Patient details modified to preserve patient identity)