RAAS blockade combined with volume restriction is the therapy of choice for most CKD patients, but the anti-proteinuric response to this regimen is often incomplete 
. Therefore it is crucial to identify interventions that can serve as (adjunct) anti-proteinuric therapies. The main finding of the current study is that the anti-proteinuric effect of the NSAID indomethacin is accompanied by reduced urinary 24 h -excretion of glomerular and tubular damage markers. The 24 h -urinary excretion of inflammation markers, on the other hand, was not affected by NSAID treatment, and the fractional excretion of MCP-1 was even increased.
The reduced urinary excretion of glomerular damage markers total IgG and IgG4 is in line with the known anti-proteinuric effect of indomethacin, and may be related to prostaglandin inhibition affecting renal hemodynamics (i.e. vasoconstriction of the afferent arterioles). Furthermore, both 24 h -urinary excretion and fractional excretion of the proximal tubular damage markers were reduced following treatment with an NSAID, and the distal tubular marker H-FABP showed a trend to reduction. This suggests that NSAIDs may have protective effects, rather than toxic effects, on the tubular epithelium. Tubulotoxic effects may be seen in rare cases, but in our study on average signs of reduction of tubular damage are observed, probably secondary to the anti-proteinuric effect of NSAID treatment. This is in concert with previously observed tubulo-protective effects of RAAS-related anti-proteinuric therapy in both clinical cohorts 
as well as preclinical studies, where also an association with protection against structural renal damage was observed 
. In addition, direct protective (e.g. anti-apoptotic) effects as implicated by in vitro studies 
may play a role. NSAID treatment did not significantly affect the distal tubular marker H-FABP, although a trend to reduction was observed. Much less is known about the effects of proteinuria on distal as compared to proximal tubular cells, although the distal tubule is considered less sensitive to the toxic effects of proteinuria. We recently reported that H-FABP is associated with albuminuria in diabetic subjects; moreover, H-FABP was the only damage marker from the investigated marker set that was significantly associated with eGFR after adjustment for albuminuria 
. Future studies should address the relevance of H-FABP modulation by antiproteinuric therapy.
Surprisingly, NSAID treatment had no effect on the 24 h-urinary excretion of inflammation markers MCP-1 or NGAL. In contrast, fractional excretion of MCP-1 even increased following NSAID treatment. This was unexpected given the anti-inflammatory properties of this drug. On the other hand, the dissociation of renoprotective effects from anti-inflammatory effects of NSAIDs is in line with preclinical studies 
. Furthermore, we recently found that in a cohort of CKD patients receiving an angiotensin receptor blocker in addition to an angiotensin converting enzyme inhibitor, urinary excretion of proximal tubular markers was reduced, but inflammatory markers did not change (data submitted
). Our finding that the extent of pre-treatment renal inflammation is independently associated with the anti-proteinuric response is also in line with preclinical studies 
. This finding may have two important implications. First, these data suggest that the renoprotective potential of NSAIDs may be limited by the absence of intra-renal anti-inflammatory effects (in spite of systemic anti-inflammatory effects as reflected by lower plasma MCP-1 and B2M levels). Second, these data suggest that the antiproteinuric, and potentially also the renoprotective effect of NSAIDs is limited to disease states with less intra-renal inflammation, for instance excluding the use of these drugs in immune-mediated renal disease.
Treatment with indomethacin induced a trend towards decreased creatinine clearance in our study, and although this reduction was relatively small, our study lacks power to exclude an effect on renal function. Previous studies demonstrated an anti-proteinuric response in the presence of a significant decline of GFR following NSAID treatment 
. The limited decline in creatinine clearance in our study may also be explained by the sodium replete state of our population. Although dietary salt restriction is part of current best available renoprotective therapy, in clinical practice it is notoriously difficult to implement a persistent reduction in dietary salt intake, and in fact in most published studies, the salt intake of CKD patients is similar to that in the general population 
. In line, several landmark randomized controlled trials published over the last years have also not been performed against the background of dietary sodium restriction. In the setting of NSAID treatment, however, sodium restriction may increase the risk of a fall in GFR 
. Of note, the antiproteinuric effect of NSAIDs has also been described to be blunted by sodium repletion at least in severely nephrotic patients 
. The current anti-proteinuric efficacy of NSAID treatment despite high sodium intake is therefore remarkable, suggesting an optimal balance between efficacy and safety of NSAIDs under liberal sodium intake. Our findings also suggest that NSAID treatment increased charge selectivity, which is known to be reduced in nephrotic-range proteinuria 
NSAIDs are nowadays rarely considered for anti-proteinuric therapy in CKD. This is for the large part explained by the availability of RAAS blockers with a more favorable potential, whereas NSAIDs may have tubulotoxic effects. Also, the reduction of creatinine clearance observed in up to 20% of patients treated with NSAIDs 
, although reversible and presumably reflecting a beneficial renal hemodynamic effect (lowering of intraglomerular pressure), has raised concern against using these drugs in patients with CKD. In our study, none of the 24 h-excretions of renal damage markers increased following NSAID treatment as compared to the period without anti-proteinuric treatment. The decrease in creatinine clearance was small and did not reach statistical significance, perhaps due to the sodium-replete condition of our patients. The limited follow-up of our study does not allow to address longer term safety of NSAID therapy.
Our study has its limitations such as limited sample size and study duration, and the post-hoc design. Given the heterogeneity of the population with respect to the underlying type of renal disease, it may be surprising that significant reductions of glomerular and proximal tubular biomarkers were found. This suggests that irrespective of the underlying disease, one or more shared pathophysiologic pathways (including proteinuria) are reduced by NSAID treatment. The current findings cannot be extrapolated to patients with more severe renal disease, who may experience more side effects such as hyperkalemia or edema, or patients on background RAAS blockade. The latter category may not tolerate the combined regimen, and a previous study has questioned the additional efficacy of NSAID when give on top of maximum RAAS blockade as anti-proteinuric therapy 
. Furthermore it would be of interest to study intra-individual changes of these biomarkers over time, and their prognostic value, in a longitudinal analysis. Nevertheless our data may provide a rationale to prescribe NSAIDs in patients with glomerular disease and proteinuria in whom immunosuppressive therapy and RAAS blockade are either ineffective or contra-indicated. Randomized controlled clinical trials investigating the long-term renoprotective (i.e. protective against loss of renal function) as well as cardiovascular effects of NSAIDs (as monotherapy or on top of RAAS blockade) are, however, warranted.
In conclusion, NSAID treatment is associated with a change in urinary biomarker profile that suggests glomerulo- and tubulo-protective effects along with reduction of proteinuria. Our findings, although they cannot be extrapolated to a setting of (maximally tolerated) background RAAS blockade with optimal volume depletion, suggest that under specific conditions NSAIDs may be used to reduce proteinuria as well as glomerular and tubular damage markers in CKD. This may be relevant given the prognostic importance of tubulointerstitial damage for the progression of renal function loss 
. On the other hand no intra-renal anti-inflammatory effects of NSAID treatment could be demonstrated and renal inflammation, another important prognostic determinant of renal function loss 
, may even be increased by NSAID treatment. NSAID may thus be considered with caution as a non-hypotensive antiproteinuric therapy when RAAS blockade is not tolerated, e.g. due to hypotension. Whether NSAIDs, either as monotherapy or given on top of RAAS blockade, have the potential to modulate long term renal outcome remains to be proven.