Marginal Dialysis: Patient characteristics influencing outcomes Dr Celine Foote Staff specialist, Concord Repatriation General Hospital Post-Doctoral Research Fellow, The George Institute for Global Health
Outline Description of marginal dialysis patients Characteristics which influence their outcomes Comorbid burden Cognitive function Frailty Malnutrition Use of risk prediction tools to put these characteristics together
We are dialysing large numbers of older patients but this has levelled off in last 5yrs In 2015, 19% of pts (n=513) starting dialysis in Australia were aged 75 New patients per million population 500 400 300 200 100 0 New patients Age specific rates - Australia 2011 2012 2013 2014 2015 Age 0-19 20-44 45-64 65-74 75-84 85+ 2016 ANZDATA Annual Report, Figure 1.3 ANZDATA report 2016 New patients per million population 400 300 200 100 0 New patients Age specific rates - New Zealand 2011 2012 2013 2014 2015 Age 0-19 20-44 45-64 65-74 75-84 85+ 2016 ANZDATA Annual Report, Figure 1.3
Older patients have substantial comorbid burden Foote et al. NDT 2012
High comorbid burden especially DM but falling prevalence of other conditions 60 Diabetes status at RRT entry Australia New Zealand % of patients 40 20 0 2005 2007 2009 2011 2013 2015 2005 2007 2009 2011 2013 2015 Non-diabetic Type 1 diabetes Type 2 diabetes 2016 ANZDATA Annual Report, Figure 1.9 40 30 20 10 0 Comorbid conditions at RRT entry Australia 2005 2007 2009 2011 2013 2015 Suspected cases included 2016 ANZDATA Annual Report, Figure 1.8 % of patients Coronary Peripheral vascular Lung Cerebrovascular 40 30 20 10 0 Comorbid conditions at RRT entry New Zealand 2005 2007 2009 2011 2013 2015 Suspected cases included 2016 ANZDATA Annual Report, Figure 1.8 Coronary Peripheral vascular Lung Cerebrovascular ANZDATA report 2016
In my unit we have used dialysis to treat patients with these characteristics: Patients > 90years old 40 35 30 25 20 15 10 5 0 Strongly agree Agree Neutral Disagree Strongly disagree 35 30 25 20 15 10 5 Bed bound patients Foote et al, unpublished data 0 Strongly agree Agree Neutral Disagree Strongly disagree
In my unit we have used dialysis to treat patients with these characteristics: 40 Patients with <1 year life expectancy 35 30 25 20 15 10 5 0 strongly agree agree neutral disagree strongly disagree Foote et al, unpublished data
In my unit we have used dialysis to treat patients with these characteristics: Nursing home residents 60 50 40 30 20 10 0 Strongly agree Agree Neutral Disagree Strongly disagree Severely demented patients Foote et al, unpublished data 40 35 30 25 20 15 10 5 0 Strongly agree Agree Neutral Disagree Strongly disagree
Survival of older dialysis patients is worse than most common cancers prostate cancer 92% breast cancer 89% renal cancer 72% bowel cancer 66% dialysis in Australia aged 45-64 60% Heart failure 52% ovarian cancer 43% dialysis in Australia aged 65-74 43% aged 75-84 27% lung cancer < 14% 5 year survival data from the Cancer Council of Australia (www.cancer.org.au) and ANZDATA 2012 (www.anzdata.org.au); JAMA. 2004; 292(3):344.
How long older dialysis patients survive is affected by comorbid burden ANZDATA report 2015
Dialysis start also impacts on function, cognition and QOL of older patients Older patients have higher risk of functional 1 and cognitive loss 2 and often also have decline in QOL 3 following dialysis commencement 1. Jassal et al, NEJM 2009 2 Murray et al, Neurology 2006 3. Da Silva-Gane et al, CJASN 2012
Comorbid burden
Higher comorbid score* predicts poorer survival 0.00 0.25 0.50 0.75 1.00 0 1 2 3 or more 0 1 2 3 4 Followup (years) Comorbid score HR 95%CI P value 1 1.43 1.19-1.70 <0.001 2 1.61 1.34-1.92 <0.001 3 2.00 1.68-2.40 <0.001 * Includes coronary heart disease, peripheral vascular disease, cerebrovascular disease, chronic lung disease and diabetes Foote et al. NDT 2012
There may be no survival advantage with dialysis with increasing comorbidity Murtagh et al. NDT 2007
There may be no survival advantage with dialysis with increasing comorbidity Brown et al. CJASN, 2015
Cognitive function
Mild cognitive impairment is a common but poorly recognized problem 16-38% cognitive impairment in all dialysis pts 1 68% in those aged over 75 2 Documented in only 3% 3 Uncertainty remains as to which cognitive instrument to use 1. KurellaTamura et al. KI 2011 2. Patel et al, HKJN 2016 3. Murray et al, Neurology 2006
Mild cognitive impairment leads to independent increased mortality risk Griva et al, AJKD 2010
Frailty
Frailty is common in CKD patients Prevalence of frailty in CKD population is twice that of general geriatric outpatient community 14% 15% versus 6% 7% 1,2 Prevalence increases markedly in dialysis population and with increasing age 44% in dialysis patients aged<40 and 78% in aged>70 3 1. Shlipak et al, AJKD 2004 2. Roshanravan et al, AJKD 2012 3. Johansen et al JASN 2007
Frailty leads to independent and graded increased mortality risk Independent 2.24x increased risk of death 1 Also 1.63x risk of death or hospitalization 1 Risk is graded with each 1-point increase in frailty associated with 1.22x increased risk of death 2 1. Johansen et al JASN 2007 2. Alfaadhel et al, CJASN 2015
Frailty has substantial impact on life expectancy Swidler CJASN 2013
Malnutrition
Malnutrition predicts increased mortality risk but may be poorly characterised Malnutrition prevalence varies between 18%-75% in dialysis patients 1 Older patients are at increased risk 2 Albumin is commonly used as a surrogate Predictor of mortality in general 3 and older dialysis patients 4 but some inconsistency in literature 5 Confounded by inflammation and comorbidity 6 1. Kalantar-Zadeh et al, AJKD 2003 2. Qureshi et al, KI 1998 3. Iseki et al, KI 1993 4. Oliva et al, J Nephrol 2013 5. Chan et al, J of Renal Nutrition 2012 6. Stenvinkel et al, NDT 2002
7 point SGA is a better predictor of malnutrition and predicts mortality in pts starting dialysis 7 point Subjective Global Assessment (SGA) is a better predictor of malnutrition Not confounded by inflammation or obesity 1 SGA score is independently associated with mortality in incident dialysis pts 2 Predictive capacity in older dialysis patients 3 1. Chan et al, J of Renal Nutrition 2012 2. De Mutsert et al, Am J Clin Nutr 2009 3. Santin et al, Clin Nutr in press
Risk prediction tools
Risk tool combining comorbidity, albumin and SQ is useful in predicting prognosis in the general HD population Cohen et al et al. CJASN 2010
http://touchcalc.com/calculators/sq
Several risk tools are available for older dialysis patients but have limitations Four prediction tools derived in older dialysis pts Derived from French renal registry 1,2, US Medicare 3 and Albertan databases 4 Risk factors assessed are dictated by availability Heart failure, arrhythmia and malignancy were consistent predictors Limited assessment of cognition, functional status Lack of clinical judgment assessment like SQ Only predict short term dialysis survival (3-6 month) 1. Couchoud et al. NDT 2009 2. Couchoud et al, NDT 2015 3. Thamer et al, AJKD 2015 4. Wick et al, AJKD in press
Summary Large numbers of older patients with considerable comorbid burden Patient characteristics which influence outcomes include: Comorbid burden Cognitive function Frailty Malnutrition Use of risk prediction tools can assist in prognostication for individual patients
Suggestions for next steps. Systematic collection in all older pts approaching ESKD of presence of: Cognitive function Frailty Malnutrition Irrespective of planned treatment pathway Improve prognostication for appropriate use of resources
Thank you for your attention
Do you systematically test cognitive function on older patients approaching treatment decisions for ESKD? 1. Yes 2. No 3. Uncertain
Should a diagnosis of dementia be a contraindication to dialysis? 1. Yes 2. No 3. Uncertain
Do you formally assess for frailty in older patients approaching treatment decisions for ESKD? 1. Yes 2. No 3. Uncertain
How do you assess for frailty? 1. Fried s criteria 2. Frailty index 3. FRAIL 4. Clinical frailty scale 5. Physician perception 6. Uncertain
There are many different ways to measure frailty 1) Physical frailty 2) Frailty index 3) FRAIL 4) Clinical frailty scale
Fried et al, J Gerontol A Biol Sci Med Sci 2001 Frailty assessments 1. Fried
Frailty assessments 2. Frailty index Number of deficits in an individual /Total number of deficits measured Based on accumulation of illnesses, functional and cognitive declines, and social situations It requires answering 20 or more medical and functional-related questions.
Frailty assessments 3. FRAIL Fatigue ("Are you fatigued?") Resistance ("Can you climb one flight of stairs?") Ambulation ("Can you walk one block?") Illnesses (greater than five) Loss of weight (greater than 5 percent) Morley JE et al., J Nutr Health Aging 2012
Rockwood et al, Can Med Assoc J 2005 Frailty assessments 4. Clinical Frailty Scale
Do you formally assess for malnutrition in older patients approaching treatment decisions for ESKD? 1. Yes 2. No 3. Uncertain
How do you assess malnutrition in your elderly ESKD patients? 1. Serum albumin 2. Body mass index 3. Anthropometry 4. Subjective Global Assessment 5. Uncertain
Do we need to capture comorbid burden? Which patients do we target? Which tool to use? Who does it formally (on top of ANZDATA)? How do we utilise the comorbid information?
New comorbidity index predicts survival in older dialysis patients Kan et al. Plos One 2013
There may be no survival advantage with dialysis with increasing comorbidity SC n=54, dialysis n=17 Chandna et al. NDT 2010
Cognitive impairment Need to screen? Who to screen? What tool to screen with? How to interpret finding of cognitive impairment into decision making? Should a diagnosis of dementia be a CI to dialysis
Cognitive impairment is a common but poorly recognized problem amongst older ESKD patients Dementia prevalence was 22% in elderly nursing home patients starting dialysis 1 30-55% cognitive impairment on neuropsychological testing in ESKD patients aged over 75 2 Uncertainty remains as to which cognitive instrument to use Multiple screening cognitive assessments exist No validation against clinical diagnoses of dementia in the ESKD population 3 Many are influenced by educational level and language fluency 1. Tamura et al. NEJM 2009 2. Tamura et al. CJASN 2010 3. Tamura et al. KI 2011
Dementia leads to increase mortality and dialysis withdrawal Dementia confers 1.5-1.9 independent increased risk of death 1,2 2-yr survival for pts with dementia was 24% versus 66% for pts without dementia (P < 0.001) 2 Also increased dialysis withdrawal (RR 2.01, 95% CI 1.57 2.57) 2 1.Kurella et al, NDT 2010 2.Rakowski et al, CJASN 2006
Frailty assessments Who to screen? What tool to screen with? How to interpret finding of frailty into decision making?
Frailty in dialysis patients leads to independent and graded increased mortality risk Independent 2.24x increased risk of death 1 Also found 1.63 risk of death or hospitalization 1 Each 1-point increase in frailty led to 1.22 increased risk of death 2 1. Johansen et al JASN 2007 2. Alfaadhel et al, CJASN 2015
7 point SGA 1. Weight loss 2. Dietary intake 3. GI symptoms (nausea/vomiting/diarrhoea) 4. Functional status (nutrition related) 5. Disease state affecting nutrition requirements 6. Muscle wastage 7. Fat stores 8. Oedema
Use of risk prediction tools Which tool? Use with limitations in mind Who uses these?
Surprise question (SQ)
SQ predicts increased mortality Would I be surprised if this patient died in the next year? Moss et al. CJASN 2008
Literature search and survey identified important factors and these were presented to nephrologists and advanced trainees
Respondents chose neither patient for 57.3% of the scenarios,
Cognition, patient choice and expected fall in QOL as most important attributes Furthermore looked at trade-offs: Nephrologists were willing to forgo 12 months of patient survival in order to avoid a substantial decrease in QOL following dialysis start Survival trade-off is similar to 15 months of life expectancy that patients were willing to give up to decrease their travel restrictions 1 1. Morton et al. CMAJ 2011