Newsflash: Numerous NSAIDs are Nephrotoxic

#NephJC Chat

Tuesday March 19th 9 pm Eastern

Wednesday March 20th 9 pm IST

Wednesday March 20th 8 pm GMT, 12 noon Pacific


JAMA Netw Open. 2019 Feb 1;2(2):e187896. doi: 10.1001/jamanetworkopen.2018.7896.

Association of Nonsteroidal Anti-inflammatory Drug Prescriptions With Kidney Disease Among Active Young and Middle-aged Adults.

Nelson DAMarks ESDeuster PAO'Connor FGKurina LM.

PMID: 30768191 Full Text at JAMA Network Open (open access)

It’s NephMadness time. If you haven’t already, head over to AJKD Blog, which is NephMadness Central this time of the year. 2019 brought these feight novel regions:

  1. Hypertension

  2. Hospitalist

  3. Hepatorenal

  4. PLEX

  5. Complement

  6. Exercise

  7. Volume assessment

  8. Pain

It is this last region, pain that is the most painful section for patients with CKD. All options including opiates are bad. NSAIDs in particular, which provide such sweet pain relief - are a major problem in CKD. This week we discuss some fresh evidence about the badness of NSAIDs. Dive in.


65 years have passed since the classic description of analgesic nephropathy by Spuhler and Zollinger.  Papillary necrosis and chronic interstitial nephritis - uncommonly seen now, but at its peak, about 1 in 6 patients starting dialysis had NSAIDs to blame. However, the offending agents (chiefly phenacetin, as well as the combinations with caffeine and barbiturates) were withdrawn from the market eventually, and NSAIDs as the primary cause has fallen off the top 10. But NSAIDs are great for nociceptive pain relief - and many NSAIDs remain on the market. Needless to say, acute kidney injury can occur with NSAIDs - whether as interstitial nephritis or the hemodynamic injury (usually with other concomitant factors).

We do warn patients with CKD to be wary of NSAIDs - they are considered at higher risk, which restricts our choice of analgesics - check out the discussion in NephMadness again. But how bad is the risk for folks who don’t have CKD? NSAID exposure and CKD risk in the more recent epidemiology literature is much less strong, if at all present.  A previous #NephJC discussion looked at the relative effects of celecoxib versus ibuprofen and naproxen and didn’t find much. But NSAIDs are available as over-the counter. Pain may get worse with the vagaries of age and its accompaniments, but being youthful doesn’t spare one of the attentions of the Algea. That’s the perceived gap in the literature this present study aims to address.

The Study


Study Design

Retrospective Cohort study


NSAID use: defined using the WHO as ‘defined daily doses’ (DDD). They were then categorized into no NSAIDs, 1-7, and > 7 DDD. What do those groups mean? That’s the mean DDD per month, averaged over preceding 6 months (prior to each observation). So > 7 DDD would mean more than 7 doses per month for 6 months. Now each dose itself is not 200 mg of ibuprofen. Check that out in the results. Also, the DDD refers to doses dispensed, not doses consumed. OTC NSAID use is also not captured in this study.

Study population

Active-duty US Army collected from January 1, 2011, to December 31, 2014 (the eTable 1 has details about the sources of data), and analysis conducted beginning Aug 2018. Those who had kidney disease would presumably be ineligible for enlistment. But since the ‘active-duty’ cohort includes incident and prevalent personnel, for those who were already in service prior to January 2011, diagnoses of AKI or CKD made in the first 6 months resulted in exclusion. Occurrence of AKI or CKD, or discharge from military service also meant exclusion from further outcome analysis.


AKI and CKD. Defined on basis of ICD-9CM: AKI (584.x, 586, 580.9) and CKD (581.x, 583.x, 585.x, 587)


Cox proportional hazards, including selected covariates. Exposure to NSAIDs grouped into 3 as above. Despite large size of cohort, propensity score matching was not used.

The National Heart, Lung, and Blood Institute funded this project in collaboration with the Uniformed Services University of the Health Sciences.


Of 827, 265 active Army soldiers who served during 2011-2014, a total of 764 ,228 met eligibility criteria for at least 1 of the 2 end point–specific analyses. The numbers for the AKI and CKD analyses are slightly different (based on the exclusion of people in first 6 months, as mentioned above).

Table 1 shows which NSAIDs were commonly used. It does not show the doses that were commonly used. So ibuprofen followed by naproxen accounted for just under a third of NSAIDs dispensed.

Table 1 from Nelson et al, JAMA Network Open, 2019

Table 1 from Nelson et al, JAMA Network Open, 2019

So what were the doses? “Of the 804, 471 ibuprofen prescriptions, 78% were for 800-mg tablets, and 88% allowed for 3 or more daily doses. Of the 376, 078 naproxen prescriptions, 96% were for 500-mg or stronger tablets, and 94% allowed at least twice-daily use.”

So the > 7 DDD per month group may mean 2400 mg ibuprofen for more than 7 days per month for 6 months for a total of at least 100,800 mg ibuprofen over 6 months. Hmmm.

Table 2 from Nelson et al, JAMA Network Open, 2019

Table 2 from Nelson et al, JAMA Network Open, 2019

Table 2 is the table 1. Mostly young men as expected. Individuals who consumed NSAIDs were different from individuals who didn’t. More men, more diabetes, more hypertension in the NSAID use categories. Highest NSAID use in the overweight (by BMI) category and then drops off - but note BMI in this cohort likely to reflect muscle mass as much as fat. Very interestingly, age went the other way: younger men had greater NSAID exposure (as it was measured here) than older men.

In the supplement, this is also reflected in the years of service and military grade:

eTable 2 from Nelson et al, JAMA Network Open, 2019

eTable 2 from Nelson et al, JAMA Network Open, 2019

The baseline characteristics is somewhat sparse on what we usually expect to see, but remember, these are (mostly) healthy young men.

So what did they find?

Table 3 from Nelson et al, JAMA Network Open, 2019

Table 3 from Nelson et al, JAMA Network Open, 2019

After adjustment for the covariates mentioned using Cox proportional hazards, the highest NSAID category (ie > 7 DDD per month) was associated with a higher risk of AKI and CKD (defined based on ICD-9 CM codes).

So did all the other covariates - at somewhat higher rates for most, eg diabetes, hypertension. And despite lower NSAID exposure, increasing age was associated with a higher risk of AKI and CKD.

Interestingly, the military rank or years in service did not have as much of an effect:

eTable 3 from Nelson et al, JAMA Network Open, 2019

eTable 3 from Nelson et al, JAMA Network Open, 2019


To sum up, in these (mostly) young men, exposure to NSAIDs in the highest category was associated with a higher risk of AKI and CKD. It is a very modest risk, but as the authors point out it works out to ‘17.6 and 30.0 additional cases per exposed 100 000 persons per year for AKI and CKD, respectively.

For the glass half-full folks, exposure to NSAIDs at lower doses (1 - 7 DDD per month) was not associated with such a risk.


Before we jump in and proclaim ‘No NSAIDs for you’ lets consider

  • all the covariates were different in table 2. People who take these NSAIDs are different than those who didn’t. Huge scope for residual confounding.

  • The dose makes the poison. The highest NSAID category reflects a pretty high NSAID use. 100,800 mg of ibuprofen over 6 months is more than what the casual NSAID popper would be taking.

  • And again, the dose makes the poison. Cumulative dose matters - not just the average dose over 6 months. It is somewhat perplexing why cumulative dose was not used in any analysis?

  • And, on dose again for the last time: the category of > 7 DDD could mean a low of, say, 7 doses of 200 mg ibuprofen = 8,400 mg, all the way to 800 mg 3 x day = 100,800 mg (7 days a month) - 432,000 (if taken daily for 6 months). That’s a huge spread to group. Can one say dichotomania?

    Lastly: it’s quite possible NSAIDs work as a cofactor in some scenario: perhaps dehydration, perhaps rhabdomyolysis, perhaps trauma & hypovolemia?  

Another small study did not show such an effect on GFR with > 2,500 pill NSAID consumption over 14 years - which would probably fall within the 1 - 7 category. Whether heavy NSAID consumption causes AKI and CKD or not, if one needs to consume 100 grams or more of ibuprofen every 6 months, it would be prudent to get those nephrons checked out.

Summary by Swapnil Hiremath, Nephrologist, Ottawa