Economic Evaluation of Frequent Home Nocturnal Hemodialysis Based on a Randomized Controlled Trial
Scott Klarenbach, Marcello Tonelli, Robert Pauly, Michael Walsh, Bruce Culleton, Helen So, Brenda Hemmelgarn, Braden Manns
Provider and patient enthusiasm for frequent home nocturnal hemodialysis (FHNHD) has been renewed; however, the cost-effectiveness of this technique is unknown. We performed a cost-utility analysis of FHNHD compared with conventional hemodialysis (CvHD; 4 hours three times per week) from a health payer perspective over a lifetime horizon using patient information from the Alberta NHD randomized controlled trial. Costs, including training costs, were obtained using microcosting and administrative data (CAN$2012). We determined the incremental cost per quality-adjusted life year (QALY) gained. Robustness was assessed using scenario, sensitivity, and probabilistic sensitivity analyses. Compared with CvHD (61% in-center, 14% satellite, and 25% home dialysis), FHNHD led to incremental cost savings (-$6700) and an additional 0.38 QALYs. In sensitivity analyses, when the annual probability of technique failure with FHNHD increased from 7.6% (reference case) to ≥19%, FHNHD became unattractive (>$75,000/QALY). The cost/QALY gained became $13,000 if average training time for FHNHD increased from 3.7 to 6 weeks. In scenarios with alternate comparator modalities, FHNHD remained dominant compared with in-center CvHD; cost/QALYs gained were $18,500, $198,000, and $423,000 compared with satellite CvHD, home CvHD, and peritoneal dialysis, respectively. In summary, FHNHD is attractive compared with in-center CvHD in this cohort. However, the attractiveness of FHNHD varies by technique failure rate, training time, and dialysis modalities from which patients are drawn, and these variables should be considered when establishing FHNHD programs.
There is a trend and fashio to promote frequent nocturnal HD based on the FHN trial implying benefit. This is of concern as the FHN trial failed to shwo survival advantage, but showed a benefit on left ventricular hypertrophy; as combined endpoints were evaluated it served to the fall impression that the combined endpoint of survival and left ventricular hypertrophy were imrpoved by Frequent HD, which wasnt the case...
Also, Frequent HD was associated with longer time, UF, as well as higher weekly KT/V in the treated group compared to standard thrice weekly HD; so to attribute the benefit to increased frequency is a misleading assertion as it could as well be due to longer weekly HD or better UF...
Form such weak, short term, data to embark on changing HD practices worldwide is a folly....the paper above also remind us of cost implications.
Nephrologist sare all too keen to embrace new practices to show that they are keeping up with the literature (often without understanding it...) and fashion (something most are attracted by....),
1. Critically examining the facts and data.
2. Examining their utility and usefulness to their own patients population.
3. Without adequate comparisons to their own practices
4. Without assessing the Risk versus Benefit; in this case the risk of jeopardising vascular access by frequent cannulation/puncture.
5. Without assessing the COST versus BENEFIT as outlines in the paper by the Alberta group. A cost benefit analysis that takes the primary intervention cost advantage over conventional HD and showing some advantages but than examining the true cost of any new intervention including the cost of complications and related medical care, the cost of technique failure and loss of vascular access, etc...ONLY THEN DOES THE TRUE COST EFFECTIVENESS OF A NEW TECHNIQUE starts to fade away as gain per QALY exceeds $50,000, the threshold for cost effectiveness in high economines.
Cost effectiveness in middle and low economies could be a gaing per QALY of as little as <$5,000, cost per QALY gained higher tha $5,000 per annum is an unacceptable cost in many low economies.
It would be great that cost benefit analysis also give some insights and figures for cost effectiveness not only in their societies and wealthy healthcare systems but also in poorer economies based on a GDP adjustement.
This would be a great excercise that would enhance the value and UTILITY of such analysis beyond a restricted healthcare system, that in which the reporting autors work within.
Prof Pierre Delanaye wrote a BLOG in OLA French:
Une étude purement observationnelle mais intéressante car spécifique aux patients insuffisants rénaux (même si la définition retenue pose un peu problème, à savoir une GFR estimée inférieure à 60 ml/min et/ou une protéinurie). Les auteurs ont comparé la prévalence du contrôle de l'HTA définie sur base d'une mesure en consultation inférieure à 140/90 ou à 130/80 ou sur base d'un monitoring de 24h avec une définition à 130/80 ou 120/75. L'hypertension blouse blanche est définie comme une HTA plus importante en consultation et une HTA masquée si c'est l'inverse (TA contrôlée en consultation mais pas sur holter de 24h). 14382 sujets espagnols ont été recrutés. 8689 n'avaient pas d'IR 72,6% étaient traités (la moitié avec plus de deux médicaments). L'utilisation et le nombre d'antihypertenseurs augmentent avec la progression de l'IR ce qui n'est pas trop étonnant. 21.7% des patients avec IRC ont une TA controlée en consultation (140/90) mais 43.5% ont un holter avec cible inf à 130/80. Le controle de la TA sur holter de 24 h ne semble pas trop différent selon le stade d'IR sur pour le controle de l'HTA nocturne qui augmente avec l'IR
14.7% des sujets sont controlés et sont concordants entre les 2 techniques
49.5% sont concordants pour une marque de contrôle de la TA
Une HTA de la blouse blanche est présente dans 28.8%
Une HTA masquée est présente dans 7% des cas
36% des sujets avec une HTA en consultation présente en fait un effet blouse blanche et 32% des patients controllés en consultation ont en fait une HTA sur le monitoring
=> intéret du monitoring en cas d'IRC?
Nephrol Dial Transplant. 2013 Oct;28(10):2553-69. doi: 10.1093/ndt/gft214. Epub 2013 Jun 4.
Cost of peritoneal dialysis and haemodialysis across the world.
International Renal Research Institute Vicenza (IRRIV), San Bortolo Hospital, Vicenza, Italy.
Peritoneal dialysis (PD) as a modality is underutilized in most parts of the world today despite several advantages including the possibility of it being offered in the remotest of locations and being significantly more affordable than haemodialysis (HD) in most cases. In this article, we will compare the cost of HD and PD in several countries to demonstrate that PD is less than, or at least as expensive as, HD. A thorough literature survey of EMBASE and PUBMED was conducted; 78 articles which compared the annual PD and annual HD costs were finally selected. Careful attention was paid to the methodology followed by each study and the year it was published in. Our final calculations included 46 countries (20 developed and 26 developing). We found that the cost of HD was between 1.25 and 2.35 times the cost of PD in 22 countries (17 developed and 5 developing), between 0.90 and 1.25 times the cost of PD in 15 countries (2 developed and 13 developing), and between 0.22 and 0.90 times the cost of PD in 9 countries (1 developed and 8 developing). From our analysis, it is evident that most developed countries can provide PD at a lesser expense to the healthcare system than HD. The evidence on developing countries is more mixed, but in most cases PD can be provided at a similar cost where economies of scale have been achieved, either by local production or by low import duties on PD equipment.
continuous ambulatory peritoneal dialysis (capd), economic analysis, economic impact, haemodialysis, peritoneal dialysis
This very interesting analysis of the cost of HD versus PD worldwide is very revealing. It shows that most Developed countries favor HD over PD as judged by the prevalence of the modalities in their RRT population; USA 93% HD versus 7% PD (1). The uptake of PD in Europe is marginally better but remains much lower than HD with for instance Germany having as few as 5% of its ESRD patients treated by PD! This, in spite of a higher cost for HD compared to PD; in the US for instance the annual patient cost for HD is around $ 87,500 compared to $66,750 for PD. It has been estimated that the US, that spends 18% of its GDP on healthcare, could save up to $1.1 billion over 5 years if the uptake of PD increased from 7% to 15% (2).
Yet there are obstacles to PD utilization in the West including: 1. Reimbursement policies (in France and Germany for instance reimbursement for HD is much higher than that for PD), 2. Physicians/Nephrologists preference, 3. Physicians familiarity with PD; a modality with which they may have little experience, 4. Patients’ preference, 5. Patients’ increasing age and dependency as well as isolation making thrice weekly HD not only a medical but also a social necessity.
The analysis under discussion paints a different cost analysis for PD in Developing countries where in the most deprived, low economies, PD is a more expensive RRT modality than HD! This has many root causes including: 1. the cost of importing PD solutions and related delivery system, 2. PD companies monopoly of pricing, 3. High export duties on such material, all combining to make PD more expensive, in some countries like Egypt 5 times more expensive [HDD/PD cost ratio = 0.22], than HD! Consequently, in Africa for instance, PD utilization is very low or non-existent outside of South Africa, where solutions can be manufactured locally bring the PD cost well below that of HD (3). Hong Kong has the highest PD utilization in the World (80%) due to the low cost of consumables and solutions (PD is half the cost of HD) as well as the PD first policy adopted in this country.
The way forward for low economy countries is to have PD as an economically viable option for RRT. For that they would have to reduce the cost of PD solutions (bags) cost considerably by:
- Imposing more acceptable prices and breaking the monopolies of major suppliers.
- Reducing import duties on bags (as has been done in Nepal and Malaysia).
- Considering the local production of Bags and PD solutions (South Africa and India).
- Negotiating import contracts with other developing countries that manufacture PD bags and solutions.
- Encouraging a PD first approach as promoted by Hong Kong and recently adopted by Thailand.
This may offer emerging countries with hardly any RRT due to unaffordability, the option of a potentially cheaper RRT modality compared to HD, thus also bypassing the cost of setting up HD units with the inherent infrastructure cost.
PD first may be, with renal transplantation as soon as possible may be a model to consider. Iran has increased considerably since 1995 its utilization of PD (manufactured locally) (4) as well as its rate of renal transplantation through a concerted centralized healthcare policy. In low and middle economies where ESRD is often a death sentence due to the lack of affordable RRT modalities, Governments have a duty to explore and provide affordable RRT options in order to comply with the moral ethics of the Istanbul Declaration banning organ trafficking. Unless, this is provided, human nature will prevail and organ trafficking will continue out of the despair of some and the greed of others…
- USRDS: http://www.usrds.org/adr.aspx
This comprehensive and well conducted meta-analysis confirms my long held bias that cardio-protection relies primarily on lowering BP regardless of the agent used. This was also the conclusion of a number of studies and meta-analysis including ALLHAT and the Meta-Analysis of Casas and colleagues in ther Lancet a number of years ago. http://www.ncbi.nlm.nih.gov/pubmed/16338452
We are often referred to the seminal study HOPE arguing that the cardioprotection conferred by RAS inhibitors is independent from their BP lowering effect and is a class specific effect that justifies their use above other anti-hypertensive agents in those at CV risk: http://www.ncbi.nlm.nih.gov/pubmed/10675071
This Mantra...like many in Nephrology has been repeated over and over again by Nephrologists keen to do their best for their patients and egged on by the Pharmaceutical Industry keen to do its best for its shraeholders....and the story went on for more than a quarter of a century inspite of dissending voices questioning the dissociation of the protective effect from the anti-hypertensive effect of RAAS inhibitors. I personally questionned that myth as far back as 2000: http://www.ncbi.nlm.nih.gov/pubmed/10754405
The confusion and issues are due to a number of reasons:
1. Pharma Industry hold of Key Opinion Leaders and Clinical Trialists over the last quarter of a century. I recollect conducting a study on progression comparing Lisinopril versus Placebo (with equal BP control) in the late 80s and study impact on progression: Result: NO difference. Outcome: Never Published...???!!!!
2. Lack of Critical Reading or Appraisal skills by most of those of us who read published material and Clinical Trials Reports; most dont bother to read the methodology section; abstracts or even title is often enough to embrace a new idea...treatment...or myth...and run with it!
3. Physicians commercialism; always keen to impress (and hence charge more...) patients with the latest treatments stemming from the latest publications even if they half understand their scientific implications as long as they fully understand their financial incolme generation implication.
4. MOST IMPORTANTLY....the alleged dissociation between the cardioprotective effect of RAAS inhibitors from their anti-hypertensive effects stems to a large extent from the fact that BP is RCTs (Clinical Trials) is seldom measured correctly; often if not invariably relying on causal office BP reading after a few minutes of rest...this is the least valuable, accurate or predictive method in terms of CVD outcomes as nocturnal, day:night and 24h Ambulatory BP measurement (ABPM) have proved much more reliable and predicitve.
In fact and of relevance, is the observation of a HOPE sub-study itself by Svensson et al (2001) who compared office BP and 24h ABPM and concluded: "Although, OBP is the correct comparator when comparing with previous large intervention trials and epidemiological studies, the effects on cardiovascular morbidity and mortality seen with ramipril in the HOPE study may, to a larger extent than previously ascribed, relate to effects on blood pressure patterns over the 24-hour period" .
So to conclude:
What seems to matter in the cardioprotection of CKD pateints is good BP control (KDIGO recommends <140/90mmHg) with due consideration to patients age and co-morbidities, this regardless of the class of antihypertensive agents used. Considerations should be then given to patients tolearbility, Risk versus Benefits of a given clkass of agents as well as Cost : Benefit ratio in countries where patients have to pay for their medication; a factor that greatly impacts on compliance and quality of BP control.
Finally, let us not forget that even in the US the majority of those with Hypertension are either NOT treated or POORLY CONTROLLED:
After years and even decades of total neglect, the Nephrology community is waking up to the fact that Creatinine Clearance is affected by Tubular Secretion of Creatinine (TSCr)...!!!
A recent article in the September 2013 issue of the Am J Kidney Disease by Sithoven and colleagues in Groningen, Netherlands shows that there is a good agreement in ADPKD between eGFR and measured GFR (mGFR) in individuals with normal renal function highlighting the fact that at normal eGFR range, tubular secretion (TS) Creatinine (TSCr) is relatively small and therefore doesnt confound estimation of kidney function by eGFR (MDRD or CKD EPI). They also argue that such correlation was maintained in a subgroup of patients followed-up for 3 years.
This somehow disagrees with teh observations of Rugennenti and his colleagues in Bergamo who showed poor agreement between eGFR and mGFR in estimating the rate of decline in kidney function in ADPKD and that both MDRD and CKD EPI failed to accurately detect changes in mGFR in those with ADPKD and declining renal function. This, in spite of a comparebale baseline GFR to that of the Groningen group at the onset but a significant rate of decline with time of around 8-10ml/min.
The discrepancy between the two articles highlight the fact well know to Nephrologists for the last century or so that the Tubular secretion of Creatinine is relatively small in normal GFR ranges but increases significantly as GFR declines to reach as mush as 50% of Creatinine Clearance when GFR is <20-30ml/min.
Therefore, eGFR is most likley to be useless in the evaluation of interventions aimed at impacting on the rate of GFR progression as when true GFR declines tubular secretion of creatinine increases confounding the use of serum creatinine as a parameter of progressive CKD. So whilst Spithoven et al are right in their statement that eGFR is accurate in evaluating true measured GFR in ADPKD, the difference between the studies and their resepctive and contradictory assertions may lay in the level of GFR and the rate of decline of kidney function; lower GFR and faster rate of GFR decline leads to increased tubular secretion of creatinine and consequently decreased agreement between measured and calculated GFRs.
This was also reported by Gaspari et al, 2013 in T2DM where formulated GFR underperformed and underestimated the rate of true GFR decline.
This again highlights the fact that TSCr cannot be ignored including in diabetes mellitus where changes in tubular secretion of creatinine have been reported due to the impact of diabetes on proximal tubules transporters of creatinine and changes related to impaired kidney function and metabolic control. Too many confounders for serum creatinine to be an accurate marker of glomerular filtration.
Furthermore, interventions themselves may impact on the tubular handling and secretion of creatinine and consequently making interpretation of the intervention of rate of decline of GFR impossible. Incidentally, this is the case with RAS (renin angiotensin system) inhibition that has been show both experimentally and clinically to improve tubular secretion of creatinine in diabetic nephropathy confounding ANY conclusions on the impact of RAS inhibitors on the decline in eGFR in diabetic nephropathy:
All the studies on the progression of CKD relying on measurements and changes in serum creatinine levels or the changes in formulated/estimated GFR (eGFR) are likely to be inaccurate due to the major confounding effect of Tubular Secretion of Creatinine. Not withstanding the additional confounder of the impact of any given intervention on Tubular Secretion of Creatinine. The BEAM/BEACON studies are tragic reminders of the unreliability of eGFR in interventions with agents that impact on tubular viability and function and consequently impact on serum creatinine/eGFR through mechanisms unrelated to changes in GFR.
Sadly and disappointingly, to date I am not aware of a single intervention study in CKD where GFR was measured serially to asses progression in the entire study population (with the exemption of the MDRD study in 1994 where iothalamate clearance was used throughout the study period).
Reliance on eGFR to test interventions that may slow the progression of CKD is a tragic and often wilful mistake.
Cystatin C and creatinine as markers of bleeding complications, cardiovascular events and mortality during oral anticoagulant treatment. Marcus Lind a,⁎, Jan-Håkan Jansson a, Torbjörn K. Nilsson b, Lisbeth Slunga Järvholm a, Lars Johansson a
a Department of Public Health and Clinical Medicine, Umeå University, Umeå, Sweden b Department of Clinical Chemistry, Örebro University Hospital, Örebro, Sweden Introduction: Impaired kidney function has been linked to both ischemic events as well as bleeding complica- tions in several clinical conditions. Our aim was to investigate if cystatin C, creatinine and calculated glomerular filtration rate (eGFR) were related to future risk of bleeding complications, cardiovascular events or all-cause mortality during oral anticoagulant treatment.
Materials and methods: In a cohort study, 719 patients on long-term vitamin K antagonist (VKA) treatment were followed for a mean of 4.2 years. Blood sampling was taken at inclusion and patients were followed prospectively. Cystatin C and creatinine were analysed and eGFR was calculated. All medical records were reviewed. Major bleeding events, myocardial infarctions, strokes, arterial emboli, and deaths were recorded and classified. Results: After adjustment for age, no association between cystatin C, creatinine or eGFR and major bleeding was found. Cystatin C was independently associated with cardiovascular events (hazard ratio 1.50 (95% CI: 1.27-1.77)) and all-cause mortality (hazard ratio 1.62 (95% CI: 1.38-1.90)).Creatinine was only associated with all-cause mortality, while eGFR was not associated with any of the outcomes.
Conclusions: Our findings underscore the superiority of cystatin C as a marker of cardiovascular risk compared to creatinine or eGFR. VKA-treated patients with increased cystatin C levels should be considered to be at an increased risk of cardiovascular events, and not bleeding complications.
Predictors of all cause CVD and mortality in a selected population showing that:
Serum Cystatin C predicts CVD events as well as all cause mortality.
Serum Creatinine predicts all cause mortality only.
eGFR: doesnt predict anything!
This observation made in a selected group of individuals on anticoagulation agrees with more general observations made in the general population by Astor et al last year (2012) that Cystatin C was superior to Cr-based eGFR in predicting outcomes; Heart failure, CAD, Mortality and even ESRD. http://www.ncbi.nlm.nih.gov/pubmed/22305758
The explanation is probably multifactorial including the fact that it is most likely that it is the non-renal aspects of either serum creatinine (wasting, sarcopenia, and catabolism) or Cystatin C (inflammation, obesity, smoking) that drive the association between low calculated eGFR and outcomes previously reported by a number of large community-based studies. http://www.ncbi.nlm.nih.gov/pubmed/21307840
Of interest in the Astor study, other filtration markers that have inflammatory associations such Beta2-microglobulin (b2M) and Beta Trace Protein (BTP) showed a superiority to serum creatinine and derived equations in predciting cardiovascular and mortality outcomes.
Overall, the inclusion of CystatinC and CysC related eGFR equations improves considerably the mortality risk prediction compared to Cr-based eGFR derivations and CKD classification. This was recently highlighted by the article of Shlipak et al (NEJM September 2013), although these authors remain unable to appreciate that it is serum Cystain C level that matters and not the eGFR derivation... http://www.nejm.org/doi/pdf/10.1056/NEJMoa1214234
So the question has to be asked, whether it is a low eGFR that predicts CV and all cause mortality as repeatdly and stubornly stated or whether it is the components that make up the eGFR calculation: sCreatinine and sCystatin C that truly determine outcomes?
And does it matter?
The answer is yes, it does matter as we have been indoctrinated over the last 5-10 years with the concept that low eGFR is worth detecting and justify population screening because...it predicts cardiovascular as well as all cause mortality. Further, such an assumption has been the basis of the new (2013) KDIGO CKD classification and its risk stratification.
Well, if all it takes is to go back to the good old serum creatinine (or urine creatinine for that matter) and rediscover that individuals in the lower quartiles of the serum creatinine range for their age are at increased risk of CV and all cause mortality, then we start wandering about the whole foundation of the eGFR based stratification of individuals. Low serum creatinine is a poor prognostic predictor based on wasting and sarcopenia, whilst a high serum creatinine tells us all we need to know about renal function and its progression...and for good measure we would also measure CystatinC in those we clinically deem to be at high CVD morbidity and mortality. Other biomarkers that are equally useful would be the good old C-Reactive protein. http://www.ncbi.nlm.nih.gov/pubmed/23975559
A step back to the future would take us back in the future to simple and reliable predictors of outcomes (renal and mortality) and away from a complicated eGFR based classification system. It would have the added advantage of avoidance of medicalisation of older "normal" individuals based on a creatinine based formulation that appears to be wanting in its primary function, that of predicting renal and cardiovascular outcomes....
There is a fascinating editorial on the 'Equity and Economics of Kidney Disease in Sub-Saharan Africa' by Luyckx and colleagues in this weeks Lancet, as part of their series focussing on kidney disease. The statistics bearing out the human/financial cost of ESRD are horrendous.
EXAMPLES OF FRAUD IN MEDICINE AND IN STUDIES INVOLVING RAS INHIBITION IN JAPAN:
Diovan Data Was Fabricated, Say Japanese Health Minister And University Officials
Following a long series of accusations, retractions, and the resignation of a prominent professor, it now is clear that data from a large Japanese study of valsartan (Diovan, Novartis) was fabricated. On Thursday officials at Kyoto Prefectural University of Medicine said that “had patient records been used in their entirety,” the Kyoto Heart Study “would have had a different conclusion,” reported AFB.
Excellent Editorial in the BMJ entitled: Too much Medicine; Too Little Care:
This Editorial makes excellent points about:
1. Redefining diseases such as Hypertension, Diabetes, obesity, hypercholesterolemia, osteoporosis and... CKD as to increase their prevalence...and raise the spectrum of "Epidemics"...!!! More an Epidemic in Diagnosis rather than a true Epidemic of Disease!
2. Redefining thresholds for disease definition leading to medicalisation of normal variations "normality"..."ageing"...etc...
3. Overinvestigating and consequently Overdiagnose; mostly incidentalomas....
4. The above raising concern, anxiety, cost...with little impact on Outcomes.
This Editorial suggests:
1. More Scepticisms from Physicians about changing thresholds.
2. More Scepticism amongts physicians about new definitions of disease
3. More Scepticism about new Guidelines and Recommendations
4. Better use to language and semantics of health and disease:
Use the terms “raised blood pressure” not “hypertension,”
“reduced bone thickness” not “osteoporosis,”
“reduced kidney function” not “chronic kidney disease” when talking with patients.