NephMadness 2015: Nephrology and Nutrition Region

Submit your picks! | For more on NephMadness 2015 | #NephMadness or #NutritionRegion on Twitter

At ASN Kidney Week anytime you go to a lecture that has to do with nutrition, it is standing room only. Nephrologists and everyone who takes care of kidney patients are intensely interested in what we put in our bodies. This is going to be a region to watch.

Team sodium is one of the most vilified and misunderstood teams in tournament history. How will they be remembered? High sodium intake driving increases in blood pressure or low sodium intake associated with increased mortality? As this matchup has important ramifications for global health, either could go deep in this year’s tournament. The nutritional requirements of AKI is another grudge match between long time rivals. Rounding out this region are two factors potentially modifying dialysis outcomes: low albumin and low BMI. In the former we look at the nature of low albumin inflammation versus malnutrition and in the later we dissect the epidemiologic data showing potential survival benefits of morbid obesity in patients on dialysis.

Wait, what? How can something called morbid be a survival factor?

Selection Committee member for the Nephrology and Nutrition Region:

 Friedman, AllonAllon N. Friedman, MD

Dr. Friedman is an Associate Professor of Medicine and Medical Director of the Hemodialysis unit at Indiana University School of Medicine. Dr. Friedman completed his nephrology fellowship at Tufts Medical Center in Boston, MA, and is formally trained in medicine, nephrology, and clinical nutrition. He has received funding from the NIH and other non-profit institutions to perform clinical research on the overlapping topics of nutrition and kidney disease. He is actively involved in the American Society of Nephrology and the American Association of Kidney Patients on a national level.

Meet the Competitors for the Nephrology and Nutrition Region

Low Na Intake: Mortality Risk Factor
High Na Intake: HTN Risk Factor

AKI Needs More Nutrition
AKI Needs Normal Nutrition

Obesity: ESRD Risk Factor
Obesity: ESRD Survival Factor

Albumin: Inflammatory Marker
Albumin: Nutritional Marker

Low Sodium Intake is a Risk Factor for Mortality vs High Sodium Intake is a Risk Factor for Hypertension

Sodium is the Wilt Chamberlain of nephrology boogiemen. Nothing in nephrology will ever earn both Rookie of the Year and MVP in its debut season, and the renal world will never see another player average 50 points a game for a season like Wilt the Stilt did in 1962. But when it comes to legacy, controversy, and importance, sodium is the Wilt Chamberlain of nephrology.

Sodium has been identified as a global health burden. Restricting sodium intake in order to reduce blood pressure and cardiovascular disease is a goal of just about every professional society or government health organization that walks the earth. The World Health Organization, US Department of Agriculture, NICE public health guidelines, American Heart Association, KDIGO, the CDC, and the Institute of Medicine have all recommended lower sodium intake. Despite all of those recommendations, US sodium intake has remained stubbornly elevated with no sign of dropping over the last 50 years.Given the ubiquity of the recommendations one could reasonably expect the science to be settled on the ill effects of dietary sodium, but emerging data over the last few years has kept the conclusions mired in controversy.

Low Sodium Intake is a Risk Factor for Mortality

Cross-sectional and epidemiologic data has repeatedly shown low-sodium diets to be associated with worse outcomes This was made clear when O’Donnell et al looked at sodium intake and adverse outcomes in the ONTARGET and TRANSCEND trials. Both of these trials looked at high-risk patients over the age of 55 with either established CV disease or high risk diabetes. Average 24-hour sodium excretion was 4.8 grams (208 mmol) or roughly double the recommended sodium intake for individuals. Expectedly morbidity and mortality rose as sodium excretion went up, but surprisingly, morbidity and mortality also rose as sodium excretion went down from the average. The mortality was lowest at precisely the average sodium intake.


 Data from O’Donnell et al.

The Belgians did a comprehensive evaluation of Flemish sodium habits and followed them for 8 years. Unlike just about any other study on sodium excretion, the Flemish Study on Genes, Environment, and Health Outcomes (1985-2004) and the European Project on Genes in Hypertension used honest-to-goodness 24-hour urine collections for all 3,681 participants. CV mortality was increased in the lowest tertile of sodium intake. During the follow-up, over 500 previously normotensive people developed benign hypertension. The incidence of hypertension was not influenced by baseline sodium excretion. Though interestingly, the cross-sectional analysis showed exactly what the large epidemiologic studies have shown, that increased sodium excretion was associated with increased blood pressure.

This curious association of increased CV mortality with low sodium excretion has also been found in the analysis of the NHANES 1, 2, and 3. Low sodium diets increase renin, aldosterone, and the sympathetic nervous system activity, possibly driving the increased adverse outcomes.

High Sodium Intake is a Risk Factor for Hypertension

He et al performed a Cochrane Systematic Review to determine the effect a reduction in dietary sodium (or more often urinary excretion of sodium) has on blood pressure and consistently found that even modest reductions of sodium for a month reduce blood pressure. In 22 trials of 1,990 people with hypertension, a reduction of salt excretion of 75 mmol (4.4 g) reduced blood pressure 5.39/2.82 mm Hg. A larger, 100 mmol (6 g) reduction in salt excretion lowered systolic blood pressure 10.8 mm Hg. The meta-analysis examined 2,240 normotensive individuals from 12 trials. A reduction in salt excretion excretion of 75 mmol (4.4 g) reduced blood pressure 2.4/1.0. A larger, 100 mmol (6 g) reduction in salt excretion lowered systolic blood pressure 4.4 mm Hg.

Translating these reductions in blood pressure to lives saved gives dramatic results. In the 2010 report of the Dietary Advisory Committee on the Dietary Guidelines for Americans, the authors estimated that a reduction in sodium intake of 400 mg/d would:

  • Reduce heart attacks by 20,000 to 32,000 per year
  • Reduce strokes by 13,000 to 20,000 per year
  • Save between 17,000 and 28,000 lives every year

From a financial perspective this represents a savings of between $12 and $20 billion dollars annually.

Various experimental studies have been done to prove the relationship of sodium intake to blood pressure, and ultimately to lives saved, but few were quite as devious as Hsing-Yi Chang’s study of Taiwanese nursing homes. Chang’s group secretly randomized 5 nursing home kitchens to either normal sodium chloride or a mixture of sodium and potassium chloride. Sodium intake in the control group was 5.2 g/d and 3.8 g/d in the intervention group. In total, 768 veterans were served by the kitchens with low salt and 1,213 were served by control kitchens. After an average follow-up of 31 months there was significantly lower cardiovascular death in the intervention group (1,310 deathsvs 2,140 deaths per 100,000 person-years). This represents a reduction of CV death of about 60% compared to the control group. The authors also noted less health care expenditures in the group fed in the low-salt kitchens. Of course, the improvements in outcomes could as much be due to the increased potassium intake as the decreased sodium intake.

The world’s government, medical, and professional organizations urge low-sodium diets because despite the holes, on balance low-sodium diets deliver reduced risk of hypertension, stroke, and cardiac disease.

AKI Needs More Nutrition vs AKI Needs Normal Nutrition

The AKI conference is one of the most important conferences in the Renal League. With 2 million people dying with AKI every year and AKI being the most common reason for inpatient nephrology consultation this is the conference to watch. Despite all the attention given to AKI, something as fundamental as how to feed patients with AKI is still controversial.

The two rivals in this conference are The Big Eaters (AKI needs increased nutritional support) and the Normal Nerds (keep nutrition needs steady despite the catabolic AKI). Another nutritional team, CKD, has almost no lessons for the AKI teams here. No one is advising protein restriction in AKI, even if it could curb uremia. Negative nitrogen balance is associated with mortality in observational trials of AKI.

Supporting the use of increased nutrition in AKI is a series of observational trials and underpowered interventional trials. There is no conclusive evidence on either side of the debate and the crowd must go with whatever observational data strikes their fancy.

Metabolic rate does not directly increase with AKI, however common co-morbidities like sepsis stimulate catabolism, resulting in increased energy requirements. If these patients are not provided any adequate protein nutrition (dextrose IVF only), a common situation in early in sepsis, they go into profound negative nitrogen balance. Energy and protein requirements rise as the body upregulates protein synthesis in order to synthesize acute phase proteins. Additionally, metabolic acidosis is catabolic, increasing energy demands. Use of renal replacement therapy can further increase calorie and protein demands.

Method Protein loss
Conventional RRT 6-12 grams of amino acids and 2-3 grams of protein per HD session link
CRRT 1.2-7.5 grams protein per 24 hours 6-15 grams amino acids per day link

All of this supports the notion that protein and energy intake should be increased in AKI because negative nitrogen balance and malnutrition is associated with poor outcomes. Additionally, increasing protein intake does increase the nitrogen balance. Interestingly, increased protein supplementation has not been shown to improve outcomes in modestly sized studies.

The cheerleaders for normal nutrition point to the fact that the same study that showed improved nitrogen balance with high doses of protein also showed increased need for dialysis due to increased uremia. Likewise metabolic studies have showed that increased protein intake is catabolic in and of itself and will increase PCR. Lastly, energy consumption and supply in septic and critically ill patients is difficult to estimate because wildly differing metabolic rates from day to day. Additionally procedures, intolerance and other considerations means that prescribed nutrition is rarely delivered as anticipated.

Beyond protein requirements, increased calories have also been tested. In a RCT of 30 versus 40 cal/kg/d, the investigators found no difference in nitrogen balance but the increased calorie load was associated with more hyperglycemia, greater insulin requirements, and increased triglycerides. Certainly not a compelling case for more nutrition. Go Normal Nerds!

Obesity: ESRD Risk Factor vs Obesity: ESRD Survival Factor

At no point in his career did Shaquille O’Neal’s points per game (career 24.6, highest season average 28.1 in ‘97-’98) exceed his BMI (35), but what a career he had, spanning 18 years. So could these two factors be linked, could increased BMI drive a long career, and could the same go for dialysis patients?

Obesity: ESRD Risk Factor

One large weakness in the reverse epidemiology theory is how can it just disappear after transplant. But study after study finds that obesity is no longer protective but harmful for kidney transplant recipients.

Meier-Kriesche et al looked at patient and graft survival after transplant based on the BMI at transplant, examining 52,000 transplants from 1988 to 1997. The authors found a U-shaped curve, similar to the one found for the normal population, with increased risk of death/graft loss at a BMI below 20 and ever-increasing risk as BMIs rise over 26. Similar results were seen in a surgical study that looked at both delayed graft function and non-death censored graft survival. For both outcomes, increasing BMI were harmful.

Additionally, while much of the increased BMI is due to fat, when attempts were made to look into what drove survival, patients with increased muscle mass driving the increased BMI did better than patients with fat driving the increased muscle mass. In fact the high BMI group with low muscle mass actually did worse than the normal BMI and high muscle mass (14% higher all-cause and 19% cardiovascular death). A second group looked at the same question but instead of using 24-hour urine creatinine clearance at the onset of dialysis (a potentially suspect methodology) they used dual X-ray absorptiometry to look at body composition. Not surprisingly BMI was positively correlated with both lean mass index (LMI) and fat mass index (FMI). After 54 months of follow-up they found the familiar finding of lowest mortality in the highest BMIs but they then had data which separated out fat and lean body mass:

Patients in the highest FMI tertile had the lowest risk for all-cause mortality, although it was not statistically significant (P = 0.134). The patients in the highest FMI tertile showed a significantly reduced risk for non-CVD mortality (P = 0.004).

Similar analyses were performed for LMI, although no significant univariate association was found between the LMI tertiles and the risk of death from all-cause, CVD, or non-CVD events.

So this data doesn’t seem to be consistent with the CrCl data. Another team used near infrared (NIR) interactance technology to determine the percentage body fat. They also used the Short Form 36 quality of Life Scoring System to expand the research beyond survival. And they tried to correlate the data with inflammatory markers. Interestingly, CRP and TNF-alpha concentrations were significantly higher in the lowest body fat percentage group than in the other 3 groups. (P=0.06). Quality of life scores worsened as percentage body fat went up. However the benefits of obesity still shined through with increased fat percentage being associated with better survival.

Screen Shot 2015-02-27 at 1.28.15 AM

And once again when investigators looked at weight loss they found the same concerning findings uncovered in other trials, loss of at least 1% body composition fat resulted in a 30-month mortality HR of 1.98.

Though the bulk of data seems to be be in line with obesity, the weight-loss data should be stratified for intentional versus unintentional weight loss. In total this theory is being driven by epidemiologic data and association does not indicate causation. It is time for a trial of intentional weight loss so we can get some real answers.

Obesity: ESRD Survival Factor

Everyone feels like they know health when they see it. And everyone knows that obesity is not good for you. It is common sense that if you are obese you need to lose weight. Strangely, these seemingly immutable laws break down in the topsy-turvy world of dialysis. Obesity, which is a potent risk factor for death in normal populations, becomes a survival factor in dialysis.

Screen Shot 2015-02-27 at 1.43.17 AM

How can a condition called “morbidly obesity” be a survival factor? From Kalantar-Zadeh’s analysis of weight and survival, “Both all-cause and cardiovascular mortality showed almost strictly decreasing rates across increasing BMI categories, ie, morbidly obese MHD patients had the greatest survival rates.” Reading that article you can almost feel the authors’ frustration at their inability to find an association of mortality with obesity, “Obesity, including morbid obesity, was associated with improved survival and decreased cardiovascular mortality, even after exhaustive adjustment for time-varying laboratory markers. These associations were independent of changes in BMI over time.”

Kalantar-Zadeh then looked at patients with low and high protein intake, obesity was still protective, right up to and including the morbidly obese. Probably most troubling was the data on patients who changed weight. Half the cohort maintained a stable weight, the other half of the cohort gained or lost more than 1% of their baseline weight. Gaining weight had a higher mortality than a stable weight but losing weight was the most dangerous of all. Consider that, the next time the transplant team recommends your patient lose weight to become transplant eligible. The obesity paradox persists regardless of dialysis vintage or patient age.

As perplexing as that is, it is not unique to ESRD, obesity has been described as a survival factor in congestive heart failure.

Screen Shot 2015-02-27 at 2.02.21 AM

COPD also has increased survival with increased BMI, as has rheumatoid arthritis. But the story truly takes a turn for the weird if you believe the results of Kovesdy et al, who looked at pre-dialysis CKD and found a survival advantage for obesity. And the advantage got larger the greater the BMI, with the best survival being reserved for patients with BMI over 36.7. The pattern was stronger in non-diabetic patients than diabetic patients, but the pattern was still there, even in diabetics. However other researchers have found obesity to be a potent risk factor for developing ESRD. Alan Go’s team used Kaiser Permanente data to demonstrate a strong BMI dependent risk of increased ESRD. One possible limitation of that data is, if obesity is actually a survival factor, one would expect more ESRD with obesity because fewer patients would be dying of other illnesses leaving them alive and at continued risk of ESRD. Thinner patients with higher risk of death would be less likely to survive to dialysis.

Albumin: Inflammatory Marker vs Albumin: Nutritional Marker

Albumin is the Kentucky of our tournament. While every other biochemical marker of uremia has come under the control of the medical team–either through drugs, dialysis, or surgery–albumin remains elusive. Every year Kentucky sends 2, 3, or 6 players to the NBA. You might think this would deplete their ranks, but Coach Calipari keeps cupboard full with year after year of epic recruiting and the team remains great.

Albumin is the most stubborn biochemical marker in dialysis patients. Nearly every other marker of uremia and metabolic control has been improving with advances in dialysis:

  • Dialysis dose has increased (URR from 63% to 73% from 1994 to 2004)
  • Hemoglobin has increased (up 2 g/dL from 1994 to 2004)
  • Creatinine has decreased (down 1.5 mg/dL from 1994 to 2004)
  • Bicarbonate has increased (up 2.5 mmol/L from 1994 to 2004)
  • Phosphorus has decreased (0.8 mg/dL from 1994 to 2004) and, importantly, has also narrowed its standard deviation

All of those characteristics have improved and may be driving steady improvement in dialysis survival. But one bad actor hasn’t budged in decades: albumin. And in case you think things might have gotten better since 2004, take a gander at the USRDS which shows average albumin stuck at 3.2 for incident patients, with a fixed 20% of prevalent patients with an albumin below 3.2.

What makes it especially frustrating is that of all those metabolic markers, albumin is the one  factor most associated with mortality. In Fresenius’ examination of their own data, low albumin was a more powerful predictor of mortality and hospitalization than access type, diabetes status, or age. And it wasn’t even close.

You would think that with it being such an important factor in dialysis mortality and morbidity, hypoalbuminemia would be public enemy number one at dialysis units. But instead we are left with some basic questions about the very nature of albumin. Is hypoalbuminemia an indicator of poor nutrition or an indicator of inflammation?

Albumin: Inflammatory Marker

The problem with albumin as a nutritional marker is the numerous and obvious places it falls down. There is no better model of malnutrition than starvation, yet in experimental conditions, starvation does not reliably result in hypoalbuminemia until very late. The degree of protein calorie malnutrition we see in dialysis patients is, quite simply, insufficient to cause the ubiquitous hypoalbuminemia seen in dialysis patients. Additionally, dialysis patients often experience rapid changes in albumin that are independent of changes in diet and have prolonged decreases in albumin that correlate better with increased inflammation. Inflammation simultaneously decreases albumin synthesis and increases albumin catabolism.

For example in multiple regression analysis, CRP replaces albumin as a predictor of all-cause and of CV mortality.

CKD (dialysis-dependent or not) is not alone in being a chronic disease associated with hypoalbuminemia. Other chronic diseases associated with wasting such as cancer and HIV also have prominent hypoalbuminemia. All of these conditions are unable to reduce resting energy expenditure (REE). The inability to reduce REE is likely part of the inflammatory response and drives the nutritional wasting that is a hallmark of these conditions.

A prospective study measured albumin and CRP while assessing nutrition through the subjective global assessment (SGA) and normalized protein nitrogen appearance rate (nPNA). When the authors related mortality to albumin they found the usual association of increased mortality with decreased albumin (HR of 1.47 for each 1-g/dL decrease in albumin). Strangely, adjusting for SGA had no effect on the risk and adjusting for nPNA had a tiny effect (HR, 1.45). However when the authors adjusted for inflammation, the risk was no longer significant (HR, 1.30; 95% CI, 0.95-1.78), indicating inflammation was a better proxy for hypoalbuminemia than malnutrition.

Part of the problem with the albumin and nutrition theory is that interventions that restore nutrition have a minimal impact on albumin. In an RCT that targeted 10 individual limitations on nutrition for dialysis patients that included etiologies as diverse as depression, shopping, and dentition, it was only possible to increase the albumin 0.21 g/dL (from a baseline albumin of 3.4), while the control group increased 0.06, a delta of only 0.15 g/dL.

In the most recent reviews of oral and parenteral nutritional supplements (Bossola et al, Dukkipatti et al, Sigrist et al, Kalantar-Zadeh et al) even when these techniques were able to improve albumin they have never been shown to improve survival. Some of this has been due to short treatments and underpowered trials but if these interventions are unable to improve survival, targeting low albumin is just a meaningless surrogate endpoint.

Albumin: Nutritional Marker

Conventional wisdom has long held that albumin reflects nutritional status. This is reflected in the numerous recommendations for dialysis patients to eat high-quality protein in response to a low albumin. See Davita’s patient facing site or the numerous patient education posters that decorate dialysis units, “Adequate protein nutrition is measured by serum albumin. If the albumin level is less than 3.6, the risks of dying or needing hospitalization increase several fold.” Accepting the link between albumin and nutrition begins to explain other observations, such as the omnipresent hypoalbuminemia and the low protein and calorie intake of dialysis patients, typically 30% below guidelines.

Supporting this are dietary interventions that improve albumin. Enteral supplements given during dialysis have a variable history of increasing albumin, sometimes the supplement works, and sometimes it doesn’t. A meta-analysis shows a small positive effect on albumin but nearly all the studies are small and short–a situation that invites publication bias. Likewise more exotic therapies such as intradialytic parenteral nutrition have been shown to improve albumin. Now granted, the improvement in albumin is modest, often as low as 0.2 g/dL. But the retrospective data show that seemingly modest changes in albumin can have dramatic effects on mortality; from Kalantar-Zadeh:

The sensitivity of measuring serum levels of albumin to predict outcomes in patients with CKD is high, with a granularity of as little as 2 g/L (0.2 g/dL) or less. In other words, a patient on dialysis with a baseline serum albumin concentration of 2 g/L (0.2 g/dL) above or below that of another patient with similar demographic features and comorbidities has a substantially decreased or increased risk of death, respectively.

In summary, albumin has long been the most accessible indicator of nutrition, its value can be influenced by changes in nutrition, and so it must be a nutritional marker.

 – Post written and edited by Drs. Joel Topf and Allon Friedman.


Submit your picks! | For more on NephMadness 2015 | #NephMadness on Twitter

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s