Hunger in Patients with Kidney Disease – Why it matters and what to do about it?

As a new pediatric nephrology fellow, my first weeks taking care of patients on hemodialysis were overwhelming. Each of my patients had complex problems and intricate physiology, and I initially learned to focus on the numbers – clearance, blood pressures and weight – in order to optimize therapies for each patient’s clinical disease and status. About three months into fellowship, however, I was stumped by a young girl with ESRD whose numbers just didn’t make sense. She was losing weight, admitted repeatedly for infections, and had worsening clinical and laboratory parameters which wouldn’t improve with interventions. Despite nutritional supplements, additional dialysis sessions, and medication adjustments, nothing changed. Finally, the root cause became clear – she was hungry, and her family couldn’t afford to buy food. 

Unfortunately, this isn’t a rare experience in the United States. Food insecurity is common, affecting 16.6% of American households and 7.8% of American children – and appears to be more common in those with chronic illness, such as kidney disease. Food insecurity is defined as limited or uncertain availability of nutritionally adequate and safe foods, or limited or uncertain ability to acquire such foods in socially acceptable ways. I have seen families frequently make trade-offs between buying medication versus food, or choosing inappropriate food for their children’s medical problems due to cost. Parents have told me they often go hungry in order to feed their children – providing a human face to the statistics published by the U.S. Department of Agriculture (USDA) Economic Research Service showing that half of all parents or caregivers experiencing food insecurity shield their children from having to skip meals by skipping meals themselves. 

Why should food insecurity matter to nephrologists? Food insecurity is an essential social determinant of health with significant implications, including poorer overall levels of health, chronic medical conditions, and lower levels of psychosocial and physical functioning. A recent report showed that two-thirds of families with food insecurity have to choose between paying for medical care or food within the past year.These findings in children mirror studies in adult ESRD patients where food insecurity was associated with worse nutritional status and poorer quality of life. Among adults with chronic kidney disease, food insecurity accelerates disease progression, and may increase the risk of ESRD. Patients with food insecurity often use coping strategies which are deleterious for their health, including medication under-use or non-adherence, postponing or forgoing preventive care, or relying on a low-cost diet of energy rich but nutrient-poor foods. 

While awareness of food insecurity has increased in pediatrics and screening for food insecurity is frequently part of general pediatrics care, there no similar recommendations in either pediatric or adult subspecialties. Given the implications for our patients and their diseases, I would argue that all patients with kidney disease, especially those with ESRD, should be screened for food insecurity as part of their routine nephrology and dialysis care. Physicians, dietitians, and other health providers often counsel patients with kidney disease on dietary interventions, and awareness of food security status is paramount in providing optimal care. Many interventions and recommendations (such as a low-salt diet or phosphorus restriction) may be challenging or impossible to implement for patients who are food insecure. Screening should be performed with a standardized screening instrument, with several available and validated by the AAP, AAFP and USDA. Screening should be coupled with timely and appropriate referrals to food resources – an area of intervention that requires innovation, collaboration and community-based solutions. While eliminating food insecurity may be the eventual goal, the first step is to screen patients at risk; as only once we recognize, name, and quantify a problem can we start to identify solutions. 

 Post by Michelle Starr, NSMC Intern Class 2018

Free Online Learning Module: Hemodialysis Kinetics 101 The ‘key’ to KT/V urea

Hemodialysis 101: The "key" to KT/V urea

Namrata Krishnan, Assistant Professor of Medicine at Yale has hit it out of the park. She has a fantastic online module to understand hemodialysis kinetics. This module is jam packed with clinical vignettes, well produced videos, explanatory posts, and links to primary literature. I would highly recommend this module to anyone wanting to know more about how hemodialysis works. You can all follow Namrata on Twitter.

Matt Sparks

Dialysis for dementia?

When I first began learning about Nephrology, I came across ‘dialysis dementia’, a progressive and fatal condition described in hemodialysis patients. Several studies in the 1970s implicated aluminium found in phosphorus binders and dialysate water as the cause. However, owing to modern techniques of water purification and the use of non-aluminium phosphorus binders, ‘dialysis dementia’ is now considered a rare adverse effect of dialysis, with a current estimated prevalence of 0.6–1.0%.

Nonetheless, moderate to severe cognitive impairment may affect 30–60% of patients undergoing hemodialysis (HD), and two-thirds of patients undergoing peritoneal dialysis (PD). The current pathophysiology of cognitive impairment in patients on dialysis might be mediated by traditional risk factors, such as older age, sex, diabetes mellitus, hypertension and cardiovascular disease; non-traditional factors, including hyperparathyroidism, elevated FGF-23 levels, vitamin D deficiency, anemia, malnutrition, inflammation, and oxidative stress; and dialysis-associated factors, such as adequacy, dialysis modality, hemodynamic instability during the procedure and solute shifts.

It was with interest then that I read recent research suggesting that peripheral clearance of amyloid-β (Aβ) by PD could help to reduce the amyloid plaque burden in the brain, potentially representing a new therapeutic approach for Alzheimer disease (AD). In this study, plasma Aβ levels before and immediately after PD in patients with CKD and in APP/PS1 mice (a standard animal model of AD) were measured. In both cases, plasma Aβ₄₀ and Aβ₄₂ levels were significantly reduced after dialysis. In the animal model, PD resulted in a decrease in Aβ levels in the brain interstitial fluid with reduced deposition even if plaque formation was well underway. The dialysis-treated mice showed reduced levels of hyperphosphorylated tau in the brain, suggesting a slowing of neurodegeneration along with decreased inflammation and increased microglial phagocytosis of Aβ in the brain. Attenuated cognitive decline was demonstrated by improved performance on the Y-maze and open-field tests.

According to the authors, this was a proof-of-concept study that restoration of the AD brain microenvironment and clearance of brain Aβ could be achieved by peripheral approaches. Yet how do we reconcile this promising experimental model with the high incidence of dementia in our PD patients? Although the USRDS data reports the risk of incident dementia to be lower for patients who started on PD than for those who started on HD, it still higher than the age-matched non-dialysis cohort. The tentative conclusion that we may draw from this is that vascular dementia is likely a far greater contributor to cognitive impairment in this population than AD.

In this study PD was very potent in removing Aβ from the blood in CKD patients. The authors highlight key differences in the PD procedure used in this study compared to standard practice. While CKD patients usually receive continuous ambulatory peritoneal dialysis (CAPD) or automated peritoneal dialysis (APD) with long dwell times of 8 hours or more, the AD mice received only 2 hours of dialysis per day. This suggests that CAPD may be even more effective at depleting the brain Aβ burden in AD patients. Similarly brain Aβ deposition appears to be lower in patients who receive hemodialysis. 

What are the implications of this study for us as nephrologists? Will we be dialyzing people for ‘dementia’ in the future? Or for other neurodegenerative diseases that may benefit from peripheral clearances such as Huntington disease or motor neuron disease?  More research is definitely needed and there will be side-effects that non-nephrologists may not appreciate but it could be an exciting area in the future.

Post by Dearbhla Kelly

Was it a PE or Just Another Other?

I have the distinct advantage and pleasure of having an Internist as my spouse.  Rarely a dinner goes by without a case discussion.  Recently, she brought up an all-too-common case…

One of her ‘’favorite’’ ESRD patients had died suddenly at home after receiving his in-center hemodialysis.  The patient was relatively young, notoriously non-compliant, had frequent fistula complications, but no known cardiopulmonary disorders.  She asked, “what do you think did this?” “Could it have been a PE or something?”

Depending on your narrative, ESRD can be either a pro or anti-coagulant state.   After all, they are exposed to heparin, they’ve got uremic platelet dysfunction, etc.  However, multiple studies have revealed the incidence of PE is 7-fold higher in ESRD compared to those with normal renal function (an incidence of PE near that of patients with active malignancy). In fact, having a PE with ESRD carries twice the mortality rate and hospitalization rate compared to those with normal renal function. Post-mortem studies would argue pulmonary embolism was a rare cause of death in ESRD, only 6.5% of hospital deaths in ESRD patients were attributable to thromboembolism.  Does this really make sense? Along with the high chance of inherent bias in autopsy studies, the risk factors for thrombosis in ESRD are seemingly endless; recent surgery, central catheterization, access thrombectomy, hospitalization, proteinuria, immobility, inflammatory states, obesity, autoimmune diseases and so on.  

More puzzling, is that recent studies have shown that pulmonary hypertension (PH) in ESRD, so dubbed ‘Uremic Pulmonary Hypertension’ is a significant entity.  Some report a prevalence as high as 50% by echocardiographic criteria, and around 10% by RHC when corrected for left sided heart failure, so called pre-capillary pulmonary hypertension.  Currently the prevalence of chronic thromboembolic pulmonary hypertension in ESRD is not known.  

It would follow that a pulmonary embolism on an already strained right ventricle would seem a perfect storm for classic PE presentation.  In fact, it has been previously postulated by Kumar et al that increased co-morbidity burden, especially cardiovascular complications associated with both these diagnoses, lead to diminished cardiopulmonary reserve, increasing both severity of PE and associated mortality.  Yet, severe signs and symptoms like hypoxia, chest pain, volume overload or persistent hypotension are quite possibly overlooked, and sooner attributed to under or overaggressive ultrafiltration, heart failure or coronary artery disease.  There is a scarcity data on clinical manifestations of PE in the ESRD population, predominantly case reports and case series.  Even DVT’s have been shown to present atypically with less extremity discomfort and far more upper extremity DVT’s vs lower extremity (30% vs 10.8%).

Ok, but wouldn’t sudden death point us in the right direction?  The limited and outdated post-mortem data have not supported massive PE as a common cause of sudden death, and a recent ERA-EDTA registry review suggest the proportion of deaths attributable to PE among dialysis patients was only 0.7% over 2 years, barely above the general population, at 0.5%. However, around 25% were either Sudden Death or Other, and as a recent review by Makar and Pun pointed out the etiology of sudden death in ESRD is quite commonly unknown and a significant number of cardiac arrests are not shockable.   One study in their review showing as many as 33% of all sudden deaths in HD patients were non-ventricular arrhythmias including asystole and pulseless electrical activity. Previous studies estimated that massive pulmonary embolism accounts for up to 13% of unexplained cardiac arrests, and the predominant rhythm of those arrests is pea or asystole in 95% of cases.  It begs, are we missing VTE?

It may be justifiable to look for PE more often, and in fact all the registry data in the world means little if we’re missing the diagnosis in the first place.  It’s akin to dogs chasing cars, what would we do if we caught one? Can we accept the risks of anticoagulation for sub-segmental PE, or incidental PE? Would you feel safe using a NOAC?  I suppose it depends on your narrative.

Posted by Corey Cavanaugh DO

OGY-3 Internal Medicine Resident, University of Louisville

Yale Clinical Nephrology Fellow - July 2017

Hemodialysis Simulator

I ran into a colleague from Children's Hospital over the weekend and she pointed me in the direction of this great education website that they have created for the education of students, residents and fellows. In particular, they have designed a hemodialysis simulator that allows you to run a dialysis session from start to finish. It includes all the patient parameters: labs, examination, fluid inputs and outputs. It then asks you to set up the circuit and start the dialysis. All of the usual parameters can be changed at any time.

This is a great teaching tool and I highly recommend it. It is geared towards pediatric patients - I doubt if many of our patients weigh 22kg - but the general concepts are the same. They also have a peritoneal dialysis simulator.

Renal Grand Rounds: Fevers on Dialysis - Not always an Infection

At renal grand rounds this week, I presented a case of a gentleman who presented with fevers, confusion, and lower extremity pain during dialysis. The patient would spike low grade fevers pre-HD and then fevers up to 105 post-HD. He had a tunneled HD line, but blood cultures were negative, and his fevers persisted in spite of changing the line. We were initially concerned for a dialyzer membrane reaction, but the time course of fevers was not consistent with either type A or type B reaction, and his symptoms persisted even after switching to an Exceltra membrane. The patient was worked up further, and his serum electrophoresis revealed 2 M components, serum free light chains showed an elevated Kappa/Lambda ratio, and he had a positive urine Bence Jones protein. His CH50 and C4 levels were undetectable, but C3 was only mildly low. Cryocrit was sent, and was positive for a type 2 cryoprotein with a predominant IgM Kappa component.

It was unclear why the symptoms of cryoglobulinemia worsened with dialysis; it was hypothesized that hemoconcentration with ultrafiltration, along with exposure of blood to cooler temperatures within the dialysis tubing led to transient complement consumption and an inflammatory reaction. The symptoms of mixed cryoglobulinemia are typically nonspecific, and patients usually present with arthralgias, fatigue, palpable purpura, and peripheral neuropathy. C4 and total complement are usually dramatically low, as in this case.

Treatment of cryoglobulinemia usually involves the use of plasmapheresis to remove circulating cryoglobulins. Steroids are suppressive in some patients, and rituximab quells formation of new cryoglobulins.  There are no studies aside from case reports about the use of eculizumab for cryoglobulinemia. Trendelenburg et al analyzed the role of complement in glomerular inflammation using mice models, and showed that mice deficient in C5 had reduced glomerular infiltration by neutrophils. Eculizumab inhibits the conversion of C5a to C5b and subsequent formation of the membrane attack complex; it therefore be theoretically useful in treating cryoglobulinemia, which causes complement mediated renal failure.

The patient was treated with 2 doses of eculizumab and then rituximab for cryoglobulinemia, and is now doing well and tolerating dialysis.

Posted by Shruti Gupta, Renal Fellow MGH/BWH

Green dialysis: what to do with discarded water used during dialysis?

Global warming is a growing threat to our health. Dialysis, while sustaining the lives of millions worldwide, has a significant environmental impact. Think of your last dialysis order. Perhaps the Kd was 500 ml/min for four hours. That’s 120 L. Dialysis units may waste up to 2/3 of the reverse osmosis (RO) water used for making dialysis. So add another 240 L. If you include additional saline used for used for priming and cleaning HD machines we may use between 400-500 L of water per 4-hour treatment. And that’s with a modest Kd of 500 ml/min.

Considering the global dialysis population, over 150 billion L of water are likely used every year with potentially 2/3 of that water discarded prior to dialysis. That’s about 40,000 Olympic swimming pools filled with wasted water. It should be noted that depending on your unit’s specific water system there may be less waste and increased recycling opportunities, but overall this remains a significant area of impact for dialysis on the world. And these concerns do not include the power and plastic needs generated by dialysis, which are also significant.

Does it have to be this way? RO “reject” water is more than just potable. It went through particulate filters, carbon filters, and water softeners prior to reverse osmosis. What emerges is practically expensive mineral water. It passes the WHO/EPA requirements for drinking water though is not legally considered potable.

Instead of going down the drain, RO reject water could be used for:

• Steam generation for hospital sterilization 
• Laundry services 
• Sanitation services 
• Landscaping 
• Almost anything you already use water for 

Reusing dialysate is more challenging given the higher conductivity (salinity) and the requirement for either recurrent RO or sorbent therapy. This can be cost-prohibitive and technologically challenging.

How much RO water does your unit’s system reject?  
Is there a way to use that water if it is not already being used? 
Beyond that, can recycling be further emphasized for both in-center and home patients? 
Perhaps in some places solar-assisted hemodialysis is an option?

One day perhaps rejected RO water can grow fruit and veg on the roofs of dialysis units. Providing sustainable, affordable nutrition to those who need it most.

If you are intrigued, the following resources offer a starting point to making dialysis more environmentally sustainable.

• UK Center for Sustainable Healthcare – Green Nephrology Network.
• Green Dialysis – An Australian based resource website.
• Green dialysis: the environmental challenges ahead.
• Reusing and recycling dialysis reverse osmosis system reject water.
• Toward greener dialysis: a case study to illustrate and encourage the salvage of reject water

Robert Rope, Nephrology Fellow- Stanford
 

How would you improve dialysis care if you had unlimited resources?

What if we could do more frequent dialysis? Would this improve QOL as potentially seen in the original FHN trial or even improve survival as seen in recent longer-term FHN follow-up?

What if we had more resources to support home dialysis? Would increasing nursing and other provider home visits and patient support make the promise of home therapy a reality for more patients?

What if we had more nurses, technicians, dieticians, or social workers in-center? Would you train them to lead intra-dialytic exercises? Give them skills in cognitive behavioral therapy? Teach patients how to eat better or how to cook? What about iPADs with educational or other productive resources for patients sitting idle in their chairs.

CJASN published an article this month calling for studies focusing on how to improve quality of life for our patients rather than debating which dialysis modality might extend life a few months/years longer. They reference an interesting qualitative study of interviews with over 30 Canadian patients, caregivers, and providers investigating potential research areas. The “top 10 research uncertainties” included items such as: enhancing communication between providers and patients; comparing dialysis modalities’ effects of QOL and mortality; addressing symptom (such as itching or fatigue) control and the psychosocial impact of ESRD; and addressing vascular access concerns. The focus on QOL by patients and their caregivers is notable and something that we should embrace.

Dialysis services in the US have an interesting, and it seems rare, quirk. 90% of patients are prescribed a therapy (in-center hemodialysis) that the vast majority of their doctors would not want for themselves. If you haven’t polled your colleagues yet you’ll likely discover they would favor PD or home hemo should they need dialysis themselves. This is a significant gap compared to our current reality.

So what would you do? 

Robert Rope, Nephrology Fellow, Stanford

Hypertensive emergency during dialysis

We had an interesting discussion yesterday about a young patient with ESRD secondary to diabetic nephropathy who was recently initiated on hemodialysis and reported headaches, chest heaviness and shortness of breath during dialysis sessions associated with a rise in systolic blood pressures to the 200’s. She was being dialyzed for 3.5 hours and was on Clonidine 0.3mg BID, Carvedilol 24mg BID, Norvasc 10mg daily.

 Intradialytic hypertension is defined as an increase in blood pressure after dialysis initiation. It is a common condition affecting up to 15% of patients on hemodialysis. For a thorough review, look at this prior blog.

The pathophysiology is likely multifactorial:

• volume overload 
• renin-angiotensin-aldosterone over activation 
• endothelial dysfunction (rises in the vasoconstrictor endothelin-1) 
• sympathetic over activation 
• EPO-related 
• net sodium gain during dialysis (hyponatremic patients dialyzed with high sodium bath) 
• electrolyte disturbances (hypokalemia and hypercalcemia) 
• dialytic removal of antihypertensives: agents that are significantly removed during dialysis include atenolol, metoprolol, lisinopril and enalapril. ARB, labetolol, carvedilol, ramipril, hydralazine, benazepril, clonidine and hydralazine and calcium-channel blockers have little clearance during dialysis (figure above)

Management include:

• increase time of dialysis with more gentle ultrafiltration
• attempt to lower dry weight
• review antihypertensives to favor those not eliminiated on dialysis 
• consider adding ACEI/ARB
• if hyponatremia, lower sodium of dialysate

 The patient above had her dialysis time increased to 4.5 hours per session, dialysate sodium was reduced to 135 (predialysis Na was 127) and an ACEI was added to her regimen. This yielded improvement in symptoms over the next few weeks and her dry weight was further lowered about 3 kg.

@LVRiella
www.leoriella.com

Infections and intravenous iron

An interesting study appeared in my email this morning examining the association between the use of iv iron and outcomes in hemodialysis patients with systemic infections. Current guidelines from all of the major societies suggest that if patients are admitted with bacterial infections in particular, iv iron should be stopped. There are both basic and clinical data supporting this recommendation although it should be said that the evidence is weaker than one would imagine.

The biologic plausibility argument rests on the fact that high iron concentrations have two negative effects in the context of infection. First, iron appears to impair the function of both neutrophils and T-cells. Patients on dialysis with iron overload have been shown to have reduced neutrophil function and phagocytosis. Similarly, impaired PMN function has been seen when neutrophils from healthy controls are incubated with ferric compounds. However, it should be pointed out that iron is vital for normal neutrophil function and individuals with iron deficiency also have impaired function. Clearly there is a sweet spot for neutrophil function but we are uncertain what it might be.

The evidence for a clinical effect for iron infusions on increasing infection frequency and severity is based on two separate strands (well summarized here). First, multiple observational studies have shown that high ferritin levels (particularly when >1000) are associated with an increased frequency of bacterial infections. There are clear limitations here. First, this finding is not consistent across all studies (4 of 14 studies did not find this association). Second, ferritin is an acute phase reactant itself and is not necessarily always elevated due to iron overload. Third, most of these studies were done in the pre-ESA era where much of this iron overload may have been due to transfusions. Finally, all of these studies were observational. As such, it is impossible to say if the risk of infection may not have been due to other factors which also lead to a higher iron requirement/ferritin levels.

The second strand of evidence is observational studies showing that higher doses or frequency of iron infusions are associated with a higher risk of future infection. Again, all of these studies are observational and so the same criticism as above applies. Similarly, there is a lack of consistent results across all studies although it should be said that the largest studies with the most patients did find an effect. However, in the most recent large study of ~120,000 medicare patients, while there was an increase in the risk of infection with higher doses of iron, the HR for infection for high vs. low dose was only 1.05. None of these studies examined the effect of ongoing iron administration on the severity of infection in dialysis patients. One could say that they are an argument against using iron in dialysis patients (which is not realistic) rather than arguments for stopping during an infection.

Today's email is thus especially welcome. This is a study published in CJASN which examines the relationship between ongoing iron use and severity of infection in Medicare patients. In total, 22,820 individual who had received iv iron in the 14 days prior to admission with a bacterial infection were included. 10% of these patients also received the iron while they were inpatients or shortly afterwards and were included as the cases. The controls were individuals who did not receive iron (i.e the iron was stopped per guidelines). The receipt of iv iron was not associated with increased short or long term mortality, LOS or likelihood of readmission in the next 30 days.

What is the explanation for this finding? One important question is why those 10% of patients did not stop iv iron after admission despite the clear consensus suggesting that it should be done? Looking at the table of demographic and clinical characteristics, there was nothing sticking out that suggested a reason for this difference and there may have been unmeasured characteristics that explained the difference. The outcomes were hard (as they would have to be given the source of the data) but may miss some subtle differences in outcomes related to the infection itself. That said, mortality and readmission are probably the most important outcomes. Finally, again this study was observational and needs to be interpreted as such.

All that said, this is an excellent addition to the literature which raises many more questions that it answers. These are important questions - too much of what we do in nephrology is based on consensus rather than clinical trials. Perhaps this might stimulate someone to do the necessary trial to answer this question once and for all.

Gearoid McMahon

EVOLVEing past Cinacalcet

One of the regular criticisms of nephrology is the lack of good quality randomized controlled trials. This is particularly the case in the world of renal bone disease where we target surrogate endpoints without necessarily knowing for certain what the effect this has on the most important outcome - patient survival. The EVOLVE trial was once such trial which examined the role of cinacalcet in preventing all-cause and cardiovascular mortality. The results of this trial were somewhat ambiguous in that the overall results were negative but this could have been affected by the slightly older age of the treatment group (after adjustment for age, there was a benefit to cinacalcet). There was, however, a wonderful rebuttal to this particular argument in a letter to the editor published later in the NEJM by Giovanni Tripepi.

Overall, the number needed to treat to prevent a single death was 500 and this is balanced by a 62% drop-out rate in the treatment group. There was also a lower rate of parathyroidectomy in the treatment group.

Last month NDT published and ERA-EDTA position statement on the use of calcimimetics in patients on dialysis. Based on a meta-analysis (which was largely derived from the EVOLVE trial), they made the definitive statement that cinacalcet should not be used in patients on dialysis to reduce cardiovascular or all-cause mortality. Further, they pointed out that although the original trials showed that PTH could be successfully lowered using these drugs and that this could reduce the rate of parathyroidectomies, there is no evidence that this has any benefits either beyond the prevention of surgery. They called for a randomized trial of parathyroidectomy vs. cinacalcet for the treatment of intractable hyperparathyroidism with hard endpoints. I second that call.

Uremic Pruritis - Algorithm for Management

Uremic pruritus affects more than 40% of dialysis patients. Even though is not considered a major complication of ESRD, it is significantly debilitating to patients.

Prior blog discussed the potential causes of uremic pruritus, though the clear pathogenesis remains obscure. Clearly, the whole skin seems to be affected and evidence of inflammation has been documented.

Recent article on KI by Mettang and Kremer nicely reviewed the available data. The management of uremic pruritus is limited though I have summarized in the figure a potential approach.

Update from Cardiology Literature: Antithrombotic therapy in Atrial Fibrillation and CKD

In 2015 I hope to blog on articles from non-renal journals that are of interest to nephrologists. First up is cardiology and a topic that has been covered in previous RFN posts (here, here, here). The optimal management of atrial fibrillation in patients with CKD is controversial as they are at both a higher risk of stroke and higher risk of bleeding than the non-CKD population; this is particularly true of patents on dialysis. Warfarin is well established in reducing the risk of stroke in patients with atrial fibrillation but the trials excluded patients with a creatinine clearance of < 30ml/min. Thus we have had to rely on, often contradictory, observational studies to guide us in this area.

A study in the Journal of the American College of Cardiology in December is the latest to investigate the net clinical benefit (or harm) of antithrombotic therapy in these patients. It was a retrospective cohort study using nationwide Danish registries to identify all patients discharged from hospital with a diagnosis of non-valvular AF between 1997 to 2011. Out of the 154,259 patients identified; 11,128 (7.2%) had non-end stage CKD and 1,728 (1.1%) were receiving dialysis. They used the CHA2DS2-VASC score to stratify the patients into high and low/intermediate risk of stroke groups. Briefly the score is calculated by adding one point for heart failure, hypertension, diabetes, vascular disease, age 65-74 and female sex and 2 points for age over 75 and a previous stroke. A score of ≥ 2 is considered high risk.
They found that among high risk patients on dialysis, warfarin was associated with a significantly lower risk of all-cause mortality (HR 0.85, CI 0.72-0.99) and there was a non-significant trend toward a reduction in cardiovascular death and a composite end point of hospitalization or death from all stroke/all bleeding. There was no benefit of warfarin in low-intermediate risk dialysis patients; indeed there was a trend toward higher all-cause mortality (HR 1.36, CI 0.96-1.94). Analysis of a sample of the non-end stage CKD patients found 19.1% were CKD stage 1-2, 20% were CKD 3, 36.4% were CKD 4 and 24.5% were CKD 5. Warfarin was associated with significantly lower risk of all-cause mortality in both high risk (HR 0.64) and low-intermediate risk groups (HR 0.62) in patients with non-end stage CKD. One caveat, highlighted in the journal’s editorial, is that certain components of the CHA2DS2-VASC score (diabetes, hypertension and heart failure) were identified based on filled prescription data, meaning the frequency of these risk factors may have been underestimated and therefore overestimating the number of patients classified as low-intermediate risk. We should therefore interpret the mortality benefit for this group with caution.

The most recent NICE guidelines in the UK, published in June 2014, do not recommend aspirin as monotherapy for the prevention of stroke in patients with AF. This study suggests the same should apply to patients on dialysis as aspirin was not associated with a lower risk of any outcome.
Analysis of the newer anticoagulants such as Dabigatran, Rivaroxaban and Apixaban, were not included in this study. They are contraindicated in patients with ESRD as they are cleared via the kidneys and drugs levels can accumulate and precipitate bleeding though their use in this setting has increased nonetheless. A study from the U.S. out this month in Circulation found that 5.9% of anticoagulated patients with AF on dialysis are started on dabigatran or rivaroxaban and that these drugs were associated with a higher risk of hospitalisation or death from bleeding compared to warfarin.

Balancing the risks and benefits of anticoagulation in patients with AF and ESRD remains complex. The current evidence suggests that warfarin remains the best anti-thrombotic available but it also has a significant potential for harm and the decision of whether or not to start treatment needs to be an individualized patient choice.

Authored by David Baird
Royal Infirmary of Edinburgh

Hyperammonemia - When should we start dialysis?

I would like to discuss a case that I recently saw in renal consult. He was a man in his 60s with history of end stage liver disease who received a liver transplant. His hospital course was complicated by anuric ATN and liver graft failure. As a result, he was started on dialysis on post-operative day 0.  Dialysis was stopped on post-op day 2 due to recovering renal function. On post-op day 3 he became encephalopathic. His ammonia level was elevated to 337 and did not improve with conventional therapy with lactulose/rifaximin. The question was whether to start dialysis or not in spite of his recovering renal function.

Causes: The urea cycle in the liver in which ammonia gets converted to urea is responsible for excretion of waste nitrogen.  Hyperammonemia in newborns is most commonly associated with inherited disorders of amino acid and organic acid metabolism. Causes in adults include Reye’s syndrome, liver failure, sepsis especially infections with urea splitting organisms, high dose chemotherapy, drugs (salicyclates, valproate), gastrointestinal bleeding, multiple myeloma, parenteral nutrition and late onset of urea cycle defects. The latter usually presents with episodic encephalopathy precipitated by metabolic stressors like infection, anesthesia or pregnancy.

Clinical features: Hyperammonemia can be life threatening and if persistent can lead to irreversible neuronal damage. It leads to cerebral edema causing progressive encephalopathy. Respiratory alkalosis is common due to central hyperventilation. Severe hyperammonemia can also cause seizures. An MRI brain is usually consistent with hypoxic ischemic encephalopathy

When to start dialysis? There are no published guidelines for when to initiate dialysis in a patient with hyperammonemia due to urea cycle defects. It is commonly indicated if the ammonia blood level is greater than three to four times the upper limit of normal or greater than 200 micromoles/L. Continuous hemodialysis is started with higher flow rates and is the most effective treatment in rapidly reducing ammonia levels.  Even though ammonia is osmotically active, the rapid removal of ammonia is not associated with disequilibrium syndrome mainly due to two reasons: First, there is a rapid equilibration of ammonia across the cell membrane. Secondly, the total amount in the blood, even in severe hyperammonemia, is only about 200 micromoles. This contributes less than 1 mosm per liter to total osmolality and therefore, even if it were all removed at once, the change in osmolality is too small to make cause disequilibrium. Contrast this with ammonia levels in the urine which are typically in the millimolar range.

The question remains whether to start dialysis in the setting of acute severe hyperammonemia (levels > 200 micromoles/L) and encephalopathy in adults with liver failure and normal kidney function. I was not able to find any literature on it and would like to know what the practice in other institutions is?  I believe since severe hyperammonemia can lead to irreversible brain damage, dialysis should be instituted. See this previous post concerning hyperammonemia in individuals with myeloma.

Posted by Silvi Shah

Baclofen in dialysis patients: Just say no!

At our hospital, we started noticing a pattern of admissions among some of our maintenance dialysis patients. In one case, an elderly woman with DM, peripheral arterial disease, and ESRD on thrice weekly hemodialysis was sent to the emergency department from her nursing facility for altered mental status. She was normal three days prior to admission, but when her nurse found her confused and difficult to arouse, they sent her to our ED. We got consulted to continue maintenance dialysis, but we noticed that she had a fluent aphasia and marked perseveration. The rest of her exam and workup was unremarkable.  I immediately called the nursing facility and had them fax over her medication administration record. Lo and behold, three days prior she was started on baclofen 5mg three times daily for complaints of lower extremity leg pain (which was probably her claudication anyway). We diagnosed her with acute baclofen neurotoxicity.

Baclofen is an oral antispasmodic that is used to treat muscle spasticity. Chemically, it is very similar to the CNS neurotransmitter GABA and it acts as an agonist at GABAB receptors, resulting in an inhibitory effect on neurons.  Baclofen is rapidly absorbed after ingestion and 90% of the drug is excreted unchanged by glomerular filtration. Therefore, patients with CKD and ESRD are at significant risk from baclofen accumulation and toxicity. The most common complaints are non-specific: drowsiness, headache, lethargy, nausea or vomiting. However, with a severe overdose, profound CNS depression occurs with respiratory depression, bradycardia, hypotonia, areflexia, myoclonus, or seizure activity.

The importance of baclofen toxicity in patients with renal failure has only recently been described and remains unappreciated by many physicians. El-Husseini et al. compiled a nice series of 41 cases of baclofen toxicity in patients with renal insufficiency. The majority of the patients had ESRD (62.9%) and the remaining either had advanced CKD or AKI. Despite being a small, heterogeneous case series, a few trends were seen. The onset of symptoms typically came two to four days after initiating the drug, and the mean daily dose was 20mg per day.

Since baclofen is small molecule (213 Daltons) and it has a relatively low volume of distribution and low degree of protein binding, it is readily removed by hemodialysis. There are only a few of clearance studies out there, but one report found that 4 hours of HD with a high-flux membrane resulted in a clearance of 120 ml/min, equivalent to a normal GFR! Since we have been more aware of baclofen, we have discovered more cases of neurotoxicity in our ESRD patients. We treat with daily sessions of high-flux HD for at least 4 hours until symptoms disappear. Our patient returned to her baseline after 3 consecutive days of HD. The authors of the case series highlight the lack of official dosing guidelines for baclofen in patients with renal insufficiency. Thus, many providers are unaware that this drug is contraindicated in those with advanced CKD and ESRD. Renal fellows are already primed to check for excessive gabapentin doses and NSAIDS, but I would add Baclofen to the list of drugs that should not be given to any patient with advanced CKD or ESRD.

Posted by John Roberts

NephMadness 2014 Part 5 - Renal Replacement Bracket

In the Renal Replacement bracket we see all the usual players, except for convective clearance, which is not in mainstream use in chronic dialysis units in the US. It is relatively unknown this side of the Atlantic (west of!) but online hemodiafiltration (olHDF) is used in many parts of Europe. olHDF utilizes diffusive clearance and convective clearance during a regular dialysis session. To use convective clearance or hemofiltration a different machine is required and also a large volume of ultrapure water is needed. This large volume of water is made and stored centrally in a dialysis unit and then put 'online' and delivered to all the machines in a unit. There of course is a financial outlay in setting up a unit to provide olHDF but after that the cost difference is a matter of a few dollars per treatment (or euros) difference (anecdotal evidence). Our understanding of the middle molecule clearance using convective clearance would suggest olHDF should provide our patients better dialysis. Also, recent trials have suggested better outcomes vs standard hemodialysis but there was difficulty delivering high volumes of replacement fluid for each treatment session in some of these trials. See Paul’s RFN post. This team is a new treatment that can potentially improve outcomes for our dialysis patients. For this reason convective clearance reaches the final four for me.

Dialysis in Neonates

Recently, there was an article in Slate magazine that detailed the very difficult decision that had to be made by a pregnant woman who discovered at 18 weeks gestation that her baby had complete bladder outlet obstruction with resulting bilateral megaureters and likely hypoplastic lungs. One of the questions related to the issue of dialysis in a newborn. In the past, this was generally not offered because of the presumption of a very poor prognosis. However, survival in neonates with ESRD has improved and it is becoming more routine in some units to offer RRT. It seems appropriate then, that KI just published a retrospective analysis of dialysis in neonates over the last 15 years.

In total, 254 patients from 32 countries started dialysis before one month of age and about half were in the first week. The most common causes of ESRD were congenital abnormalities of the kidney and urinary tract, cystic kidneys (mostly ARPKD) and cortical necrosis. More than 90% started with PD with most of the remainder starting HD because there was a contraindication to PD (e.g. recent abdominal surgery). One patient had a transplant in the first week but died shortly afterwards. Fifteen individuals had some later recovery of renal function and were able to come off dialysis, at least temporarily.

Overall, survival was excellent - 2-year survival was 81% and 5-year survival was 76.4%. In cases where the cause of death was known, about 2/3 were due to sepsis. Interestingly, the only factor that was significantly associated with an increased risk of death was the presence of concomitant neurological disorders (HR 5.2, CI 1.7-15.4). This may suggest that there was considerable selection bias with neonates who were considered unlikely to survive due to the presence of severe co-morbidities not being commenced on dialysis.

Long term, 45 patients received a transplant in the first 2 years of life. The 5-year patient and graft survival in these individuals was 84.2%.

As expected, there were significant co-morbidities present in these children. 20% had neurodevelopmental delay and 12% had pulmonary problems, mostly hypoplasia (the kidneys are an important source of amniotic fluid which is required for proper lung development). Birth length was below the 3rd percentile in 43% and 63% had growth retardation at 2 years. About 40% were on antihypertensives at 2 years and about 86% required treatment with EPO.

Overall, survival in this very vulnerable group was much better than I would have expected. Of course, given the presence of multiple co-morbidities, it is uncertain what their longer term outcomes would be. However, it does provide some hope to a group who, not so long ago, would have been considered hopeless. Consider that in 1998, a survey of French neonatologists found that 24% would never consider RRT in a neonate whereas a more recent survey suggested that 98% of neonatologists would offer RRT in at least some patients.

Warfarin for Atrial Fibrillation in Dialysis Patients: Where’s the data?

ESRD patients have an increased incidence of atrial fibrillation (AF) and a higher risk of stroke compared to patients with normal renal function. However they also have an increased bleeding tendency due in part to uremia and the associated platelet dysfunction. Anticoagulation with warfarin for AF is widespread in the general population, where the sense of risk and benefit are better understood. Like many other conditions, ESRD patients are routinely excluded from clinical trials in AF and therefore we lack hard data on how to manage our ESRD patients with AF. We must rely on observational data, which is often contradictory. Firstly, increased INR variability has been demonstrated in dialysis patients treated with warfarin and they spend much time outside of target INR range. Some registry data supports warfarin use with a suggestion of improved survival in dialysis patients with AF while other studies demonstrate an increased risk of stroke with warfarin [here, here, here]. Overall, the majority of studies do not support a protective effect for warfarin in ESRD patients with AF.

A current study from Canada in Circulation weighs in on the issue. The authors conducted a retrospective cohort study of patients > 65 years admitted with a diagnosis of AF from 1998 to 2007. There were 1,626 dialysis patients (46% received warfarin) and >200,000 non-dialysis patients. In a multivariate analysis, warfarin use in the dialysis patients was associated with similar risk of non-hemorrhagic stroke but a significantly higher risk of bleeding (defined as intra-cerebral, intra-ocular, GI, unspecified hemorrhage & hematuria; HR 1.44). The non-dialysis patients did see a lower incidence of ischemic stroke with warfarin use.

Our ESRD patients comprise a unique cohort with labile, often high blood pressure, repetitive AV access puncture, proven variability in INR and usually anticoagulation use during dialysis. They are certainly high risk for bleeding. Moreover, warfarin use is associated with accelerated vascular calcification in CKD patients and calciphylaxis, an admittedly uncommon but devastating condition. Unfortunately we suffer from a lack of alternatives to warfarin. Accumulation of low-molecular weight heparin in ESRD precludes its use and there is no experience with newer agents such as direct thrombin and Factor Xa inhibitors.
Many authors, including those at UpToDate, recommend warfarin use with AF and an eGFR<15mls/min. They are equivocal when CHADS2=0, but this is rare in a dialysis patient. Remember CHADS2 includes congestive heart failure, hypertension, age ≥75, diabetes and previous stroke/TIA. Despite older age being part of the CHADS2 score suggesting treatment efficacy, age >75 has been associated with a particularly high risk of bleeding in dialysis patients treated with warfarin. Also, CKD is part of the HAS-BLEED score which predicts high risk of hemorrhagic complications to warfarin when the score is ≥3. Many dialysis patients would fall into this category based on hypertension, older age and concomitant meds (85% in the current study in Circulation). I often feel uncomfortable using warfarin in ESRD and certainly feel it should be an individualized patient choice. With the current evidence (or lack of it) to guide us and the significant potential for harm, withholding warfarin for many of our older dialysis patients would not seem unreasonable.

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.

eJournal Club - Session length and weight gain

This month's eJournal club concerns a paper published by a fellow from our institution that attempts to get to the bottom of an interesting question: in patients with a high interdialytic weight gain (IDWG), is the dialysis session length (DSL) or the total volume gained more important? Both of these have been associated with increased mortality and both lead to an increased ultrafiltration rate (UFR). However, because they are interrelated, it is difficult to say which is more relevant.

For this study, the authors looked at more than 14,000 patients attending dialysis in the US. They excluded patients at extremes of session lengths and those who did not gain any weight between sessions. Patients with a URR less than 65% were also excluded to rule out any effects of underdialysis. The mean IDWG and DSL over 30 days were chosen as the exposures of interest. Interestingly there was high correlation between the 30-day IDWG and DSL and the 60 and 90-day means. The outcome was death from any cause. For the purposes of the analysis, the participants were divided into 2 groups for each exposure - less than or greater than 3kg IDWG and less than or greater than 240 minutes for DSL. A matched case-control study design was used.

Not surprisingly, the patients with higher IDWG tended to be younger, male, AA, had higher blood pressure and a higher prevalence of diabetes and CHF. Patients with lower DSL were more likely to be female, older and were less likely to have diabetes, CHF and CAD.

For the DSL analysis, lower DSL was associated with a HR of 1.32 (1.03-1.69) for mortality after full adjustment. For the IDWG analysis, increased IDWG was associated with a HR of 1.29 (1.01-1.69) for mortality. Thus both DSL and IDWG were independently associated with mortality.

What does this study mean for clinicians. It suggests that targeting both of these interventions could be useful. However it should be pointed out that this is an observational study and that they could not show that changing any of these exposures changed risk. Also, the because of the study design, the authors can only state conclusively that both are associated with mortality and not which one is more important. These conclusions may seem obvious but it is important to have good evidence to present to patients who may be frustrated with our requests to increase times and reduce fluid intake.

Head over the eJournal Club to continue the discussion of this paper.