Diabetes and CKD: Pitfalls - Monitoring response to therapy

There are two important issues to remember when assessing glucose control in patients ESRD, one relating to immediate blood sugar measurement, and another relating to longer term diabetes control.

The first has previously been mentioned on this blog. The use of icodextrin in PD solutions is increasingly common as a means of increasing fluid removal with less absorption of the solute. Icodextrin is not generally metabolized in the peritoneum but small quantities can cross into the systemic circulation where it is metabolized to maltose. Some commercial blood sugar test strips are unable to differentiate between glucose and maltose in the serum. This is not normally an issue because there is very little sugar apart from glucose in the blood. However, in patients on PD, the presence of maltose can lead to falsely elevated blood sugar readings with certain analyzers. This has lead to at least one death in a patient who was inappropriately treated with insulin in this setting. Of course, icodextrin is not the only potential source of maltose - certain IG preparations can also contain maltose resulting in similar presentations.

Traditionally, long-term monitoring of glucose control is done by regularly measuring HbA1c levels. However there are some concerns regarding the use of HbA1c in dialysis patients. RBCs in patients with ESRD tend to have shorter half-lives while the use of EPO appears to affect Hb glycation in unpredictable ways also. The relationship between HbA1c and mortality in dialysis patients is complex and contradictory results have been noted in various studies. A recently published study using data from the DOPPS study found that HbA1c was associated with mortality but only at substantially higher levels that would typically be considered normal.

One possibility is to use alternative measures of glycemic control such as glycated albumin or fructosamine. These have the advantage of not being affected by the Hb concentration. However, in contrast to HbA1c, they measure short-term glucose control only - for example, glycated albumin is a marker of glucose control over a period of about 17 days. One recent study found a significant association between glycated albumin and mortality while HbA1c was not as useful. It's possible that the reduced association between HbA1c and glucose control in diabetics with ESRD may be a contributing factor in recurrent episodes of hypoglycemia in some patients.

Diabetes and CKD - Pitfalls: Estimating GFR

The routine use of estimating equations for GFR has revolutionized the way that we view renal disease over the last 15 years and although some argue that this has lead to overdiagnosis of CKD, I believe that this has been an extremely positive development both in clinical and research terms. One criticism of the MDRD equation in particular was that it did not perform well in patients with near normal GFR and the CKD-Epi equation was introduced, at least in part, because of this limitation. However, there remain concerns that in patients with diabetes, particularly in those with hyperfiltration, this formula still does not perform sufficiently well.

To answer this question researchers in Italy took patients from two clinical trials who had serial measured GFR for up to 8 years and compared the results with simultaneous estimates of GFR using the 14 different equations. Of the 600 patients included, 15% were hyperfiltering and 13% had a reduced GFR. Overall, all but one of the equations underestimated GFR in the group as a whole. The single equation that overestimated GFR (Ibrahim) tended to overestimate at all levels. The range of differences between the mGFR and eGFR was -40 to +20 ml/min/1.73m2 and the mean percent error (MPE) ranged from -28.14 to 0.98%. Not unexpectedly, the majority of the error was related to underestimation of GFR in patients with hyperfilatration (MPE -12.8 to -36.7%). It is notable that the MPE was lowest in participants with hyperfiltration using the CKD-Epi equation. In this group, the mean mGFR was 132 ml/min/1.73m2 while the mean eGFR ranged from 83-114 ml/min/1.73m2.

The bias was far lower for the normofiltration and low GFR groups. Because the authors had longitudinal data also, they were able to look at the ability of the formulas to measure GFR decline over time. Given that all of the equations underestimated GFR at baseline, it is unsurprising that there was systematic underestimation of GFR decline over time, particularly in the patients with hyperfiltration. This was less marked in the patients with CKD at baseline. Five of the equations actually estimated that GFR was increasing in the patients despite a consistent decline in mGFR.

This is all not to say that these formulas are not useful. It is always important to recognize the limitations of your tools and one of the major issues here is that creatinine is used as the marker of kidney function with all of the limitations that this introduces. It should also be said that although the agreement with mGFR might not be great, we know from large EPI studies that an eGFR of less than 60 ml/min/1.73m2 is associated with poorer outcomes and this is true no matter what the cause of the disease. The take home from this is that it is not possible to accurately diagnose hyperfiltration in diabetic patients without over nephropathy using current creatinine-based estimating equations and that other signs should be taken into account when assessing these patients.

(Click on images to enlarge)

Bardoxolone - The Final Chapter - Part 2

One of the issues surrounding the study of bardoxolone in animal models was that, because of the way that it is metabolized in rats and mice, it is highly toxic when given for long periods. This specific toxicity is not present in humans. As a result, it was not possible to study the drug directly. An alternative was to study analogues of the drug. A group from Italy have just published the results of a study of an analogue of bardoxolone, RTA 405, in rats with type 2 diabetes. This study was accepted for publication before the termination of the beacon trial but raised some important concerns which are even more salient now.

RTA 405 caused significant weight loss and elevation of transaminases in treated rats. Also, proteinuria increased threefold. When this rat model is treated with ACEi, there is normally a decrease in proteinuria and renal damage. When the rats were treated with RTA 405 and an ACEi simultaneously, there was some reduction in proteinuria but not to baseline. The proteinuria was accompanied by evidence of severe glomerular and tubular damage.

It is possible that these results are due to a metabolite of RTA 405 which may not be present in humans and as a result, the findings are not necessarily generalizable to humans. However, these adverse effects are similar to those initially reported in the NEJM. The mechanism for the increased renal injury is uncertain at this time.

Previous posts on this topic are here, here, here, and here

Image of the Month - 2

A man in his 50s with a history of diabetes, renal stones, gastric bypass surgery and CKD stage III/IV presented to the clinic with fevers, chills and vague abdominal pain. A urine culture was positive for E Coli and he was treated with a quinolone with resolution of his symptoms. In view of his previous history of renal calculi, a CT abdomen was ordered.

The CT scan revealed a 5 x 2.2 cm mass lesion in the right renal/suprarenal region inseparable from the kidney and the right adrenal gland and the differential diagnosis was either a RCC or adrenal carcinoma. He then proceeded to an MRI abdomen.

This again showed a mass in the right suprarenal region, inseparable from the right kidney and adrenal gland with some central necrosis. At this point, the decision was made to proceed with a partial nephrectomy for likely carcinoma.

This is a low-power view of the renal cortex. There is diffuse global glomerulosclerosis involving approximately 80% of glomeruli. There is also significant tubular atrophy and interstitial fibrosis with associated areas of inflammation.

There was focal perirenal and intrarenal scarring with disruption of the renal capsule.

The adrenal gland was essentially normal apart from some focal inflammation and adhesion to the capsule. There was no evidence of any malignancy and it is likely that the changes seen were a result of infection which had resolved by the time of the surgery.

Higher power view of the preserved glomeruli revealed changes characteristic of diabetic nephropathy - nodular glomerulosclerosis.

Interestingly, he also had many tubules containing oxalate crystals, likely related to his previous gastric bypass surgery. There was associated acute tubular injury.

This case lead to an interesting debate in our conference. Should he have been treated for a longer period with antibiotics and then rescanned prior to the resection. In the end, the consensus was that the treatment he received was appropriate. Multiple imaging studies were done that were suggestive of malignancy and he had constitutional symptoms including weight loss and fever. A review of all tumor nephrectomies performed at BWH a few years ago revealed that 1/110 cases was not actually tumor. It is hard to argue in that setting that delaying the resection is the appropriate management. Of course, in this case, there was the added complication of advanced CKD and he has now lost some of his residual GFR (although his creatinine has returned to the pre-surgery baseline).

One final image: on the MRI scan, he was incidentally found to have multiple gallstones - I just thought that the picture looked really cool.

Click on any image to enlarge

Diabetic Nephropathy, or not?

A man in his 30s with a history of type 1 DM and chronic hypokalemia was referred to the renal clinic for investigation of CKD. His creatinine was 1.8g mg/dl.  His DM was well controlled without any evidence of retinopathy.  Urinalysis did not show any proteinuria or hematuria.  His renal biopsy showed focal tubular atrophy, dystrophic calcification in the scattered tubules, and did not have any signs of diabetic nephropathy.  His renal biopsy findings were therefore attributed to chronic hypokalemia.

Hypokalemia can cause kidney damage if it persists for longer than one month.   Chronic hypokalemia can cause non-specific vacuolar lesions in the epithelial vessels in the proximal tubules.  Typical renal biopsy will show interstitial nephritis, fibrosis, tubular atrophy and cyst formation.  The pathogenesis of hypokalemic nephropathy is not clear.  The hypotheses are:  1) complement activation and tubular cell damage by hypokalemia induced renal ammonium production 2) stimulation of cell growth and proliferation by intracellular acidosis 3) increased production of growth factors (VEGF, IGF-1) and cytokines by hypokalemia through an uncertain mechanism.

After further work-up, our patient was diagnosed with Giltelman syndrome.  He was started on potassium replacement and his Cr has remained stable since then.  

Posted by Jie Cui

Bardoxolone - Part 3

Last year, we had a post about Bardoxolone for the treatment of diabetic nephropathy. After one year of treatment, eGFR increased significantly in patients treated with Bardoxolone relative to controls. At the time, significant concerns were raised about the fact that albuminuria also increased in patients receiving the drug and it was uncertain both what the mechanism of this was and whether there would be any deleterious consequences. 

This week, an article was published in JASN which goes some way towards explaining the reason for the proteinuria. Under normal circumstances, a significant quantity of albumin is filtered in the glomerulus. Almost all of this albumin is reabsorbed in the proximal tubule by the cubilin-megalin complex. Defects in cubilin and megalin have been associated with albuminuria in animals and humans. It turns out that Bardoxolone downregulates the expression of megalin in monkeys in the proximal tubule. Thus, the increase in albuminuria may be due to decreased effectiveness of the retrieval process. At one year, there were no significant differences in renal histology between the treated monkeys and controls. 

This is a fascinating finding. There has been a lot of work done in the last few years regarding the effect of albuminuria itself on renal fibrosis. Even in the absence of any vascular changes, overload albuminuria is associated with increased fibrosis in animal models. So, why is there no damage seen in these monkeys? One of the postulated mechanisms of albuminuria-induced fibrosis is that megalin itself acts as a transmembrane receptor and stimulates EGF production in the presence of excess tubular albumin. This ultimately leads to increased interstitial fibrosis. The loss of megalin in treated monkeys means that this pathway might be downregulated. It should be said that these were healthy monkeys and the effect in humans with more albuminuria at baseline could be different. That said, this study goes some way towards alleviating some of the concerns that were raised last year.

To biopsy or not to biopsy

Diabetes is a common cause of CKD and the prevalence of diabetic nephropathy is increasing. A question that sometimes arises in the clinic is when it is appropriate to biopsy a patient with a presumptive diagnosis of diabetic nephropathy and are there any signs that may suggest that there is an alternative diagnosis. It is reassuring that recent data have suggested that renal biopsies in low-risk patients are associated with a very low risk of adverse consequences but I still don't think that anyone would recommend biopsying all patients who present with diabetes and CKD.

So how do you decide who to biopsy. It has been suggested in the past that the presence of hematuria was associated with a lower incidence of true diabetic nephropathy while diabetic retinopathy suggested the opposite. A study was recently published in the Journal of Diabetes Investigation which looked at the results of renal biopsies in patients with diabetes. In total, 55 patients with T2DM were included in the study. These were not random patients with diabetes; all of them had a clinical course which suggested that an alternative diagnosis might be present. 30 of the patients had true DN while 25 had no evidence of diabetes. Of these 13 had IgA nephropathy. One patient had a crescentic GN. The duration of diabetes was not a good predictor of the likelihood of DN. This is not altogether surprising when you consider that this is duration since diagnosis and the patients may have had diabetes for signficantly longer without being aware. Poor glycemic control and the presence of diabetic retinopathy were significantly associated with a higher likelihood of DN. No patient in the non-DN group had diabetic retinopathy which again goes to show the usefulness of this examination in stratifying patients with presumptive diabetic renal disease. This goes to show that the advice that we have been getting in the past is good - the absence of diabetic retinopathy should make you suspect an alternative diagnosis.

One interpretation of this study is that patients with atypical presentations and diabetes should have a biopsy because there is a good chance of an alternative diagnosis. Another, more conservative interpretation is that even in patients where there is a high pre-test probability of a non-diabetic lesion, the majority of patients will have diabetic nephropathy and in those that don't the treatment will most likely be the same in any case.