User blog: Arif Khwaja
In the UK all citizens are entitled to free care at time of need as part of the country's National Health Service (NHS). Introduced in 1945 after the war by a Labour government the NHS ensures that all people receive free healthcare irrespective of their financial status and it has been incredibly popular amongst patients and clinical staff. A cornerstone of the NHS has been the investment in General Practioners (GPs) in primary care. GPs are family doctors responsible for the overall health of their patients and have been in recent years incentivised to take on management of non-communicable chronic diseases such as CKD, Diabetes and hypertension.
Last week I came across an interesting, if somewhat depressing analysis of healthcare resources in my own city of Sheffield from Dr Parvez Hossain and colleagues. Whilst the number of GPs per head of the population was the same across the city when social deprivation was factored in there were some startling differences - those with non-communicable diseases from socially deprived areas had far fewer GPs. Or to put it another way the number of GPs per head of the population was the same for the most affluent and the most deprived -BUT the poorer you were the more likely you were to have an NCD and so you effectively had access to less healthcare. Reading the paper reminded me of the graph I saw at medical school on the impact of vaccination and antibiotics on the decline of infectious diseases such as TB, cholera and typhoid - the effect of all these interventions was minimal when set against improvements in housing and sanitation. Similarly the battle against NCDs such as diabetes and CKD will require societal and governmental action to have the biggest impact.
The paper highlights what has long been recognised as the inverse care law of healthcare - ie those that that need healthcare the most often have the least access to it - even in systems such as the UK which have a relatively sophisticated and equitable form of healthcare. As individual clinicians we need this kind of data to understand the population we work in so we can try to impact on healthcare at a societal level.
An interesting pilot study is published in this months cJASN adds more data to the never ending debate about the benefits of calcium binders versus non calcium based binders. A multicentre RCT based in Napoli, Italy assigned 212 CKD3/4 patients to sevelamer or calcium carbonate. The primary endpoint was all cause mortality and 22/105 patients died in the calcium carbonate arm compared to 12/107 in the sevelamer arm (p<0.05) over 3 years follow up. The time to dialysis inception was longer in the sevelamer arm though the criteria for starting dialysis were not standardised. One of the criticisms of the DCOR study ( a huge RCT which failed to show any difference in outcomes between sevelamer and caclcium binders in a prevalent dialysis population) was that by the time dialysis starts all the vascular calcification has taken place - therefore the authors are to be commended for doing an RCT in the predialysis population. Furthermore the de novo onset of coronary artery calcification was 81.8% in the calcium group versus 12.8% in the sevelamer group.
However there are a numbers of issues to bear in mind:
i) the study wasnt blinded - not clear why not
ii) there was not target phosphate to achieve - clinicians just had to maintain phosphate concentration between 2.7-4.6mg/dl
iii) the problem with this approach is that the final and time-average phosphate concentration was significantly lower in the sevelamer group than the calcium group! Thus its impossible to know whether the improvement in mortality is due to better phosphate control or use of a non-phosphate binder..... in fact in adjusted models for baseline and time-varying covariates the protective effect of sevelamer was lost!
iv) to my mind the most interesting part of the study is the that its the first RCT to suggest that lowering phosphate in the predialysis population may improve outcomes.
v) this study reinforces the need for an RCT in the predialysis population to address 2 specific issues - what the optimal phosphate level is (if any) and whether using a calcium vs non-calcium binder makes a difference. To be fair to the authors they acknowledge that this is a pilot study and recognise the need for a large RCT
vi) Whilst in my opinion the cost of sevelamer doesn't justify its widespread use in the predialysis population once the drug becomes generic and prices fall it would be reasonable to use sevelamer as an alternative to calcium based binders.
Perhaps one of the more interesting breakthroughs in ANCA associated disease over the last few years has been the finding that anti-lysosome associated membrane protein 2 (LAMP2) antibodies were found in patients with ANCA disease. Anti-LAMP2 antibodies could induce a crescentic GN when injected into rats and like ANCA could activate neutrophils. Furthermore it was postulated that molecular mimicry between a LAMP 2 epitope and the bacterial adhesion protein FimH ( expressed on gram negative bacteria) was the mechanism behind LAMP2 autoantibody production. Again rats immunised with FimH developed a crescentic GN and anti-LAMP2 antibodies - i.e. infection with gram negative bacteria could trigger auto-antibody production of anti-LAMP2 which in turn can induce a crescentic GN.
2 conflicting reports in this months JASN suggest further work is needed to test whether the story pans out as neatly as suggested. Roth and colleagues in found the prevalence of anti-LAMP 2 antibodies to be around only 20% in a cross-sectional cohort from North Carolina and a similar prevalence in a second cohort from Boston. Furthermore they were unable to replicate earlier work demonstrating that injection of anti-LAMP2 antibodies induced a crescentic GN in rats.
In contrast Kain and colleagues analyses samples from three European centres - Cambridge, Vienna and Groningen and found anti-LAMP2 antibodies in 81% of patients though antibody levels quickly disappeared after treatment.
A number of explanations have been postulated to explain these discordant results including methodological differences in the assay used to identify anti-LAMP2 in the two groups and the fact that 2 different patient populations were being studied. Its important to point out that the US data analysed cross-sectional samples and so many of the patients may have been in complete or partial remission. In contrast Kain and colleagues were able to analyse sera from untreated patients at presentation which may explain the higher prevalence in their study. Furthermore they were able to demonstrate that antibody levels appeared to relate to disease activity.
Clearly methodological issues surrounding the assay will be key to resolving these conflicting data and analyses from further centres will be required to ascertain with anti-LAMP2 antibodies will be a useful biomarker if disease in these patients and to clarify the role of infection as an initiating factor for the disease.
Anti–LAMP-2 Antibodies Are Not Prevalent in Patients With Antineutrophil Cytoplasmic Autoantibody Glomerulonephritis
Anti–LAMP-2 Autoantibodies in ANCA-Associated Pauci-Immune Glomerulonephritis
: . J Am Soc Nephrol : – ,
Interesting data published in this months JASN again challenges the ‘one-size fits all’ approach of many guidelines in nephrology. The Correction of Anemia and PRogression of Renal Insufficiency in Transplant patients (CAPRIT) study from Amiens France, was an open label RCT that randomised patients to receive epoietin-beta to achieve either a normal haemoglobin (13-15g/dl) or a lower haemoglobin (10.5-11.5g/dl). There was no difference in baseline characteristics of the 125 patients who were randomised to the study. At 6 and 12 months the normal haemoglobin group had improvements in quality of life scores compared to the lower haemoglobin group. At 2 years progression to ESRD and return to dialysis occurred in 3 patients (4.8%) in the normal haemoglobin group and 13 patients (21%) in the low haemoglobin group (P=0.01) The death-censored graft survival at 2 years was 94.6% in normal haemoglobin group compared to 80.0% in the low haemoglobin group (P=0.01). The primary outcome was difference in eCrCl by Cockroft-Gault. At year 2, the mean eCrcl decreased by 2.4±1.1 ml/min per 1.73 m2in the normal haemoglobin group and 5.9±1.1 ml/min per 1.73 m2in the low haemoglobin group (P=0.03). There was no increase in CV events in the normal haemoglobin group. Theese results contrast the epo studies in the CKD population where normalisation of haemoglobin was harmful – TREAT, CHOIR, CREATE.
So is this a practice changing paper? I don’t think so because the study was much, much smaller than the CKD trials with a lower prevalence of cardiovascular disease – therefore the study may simply have been underpowered to detect differences in adverse cardiovascular events. Postulated mechanisms by which epo may improve transplant outcomes include a possible immunomodulatory effect or correction of tissue hypoxia and clearly warrant further investigation. Furthermore the potential impact of epo on malignancy could not be evaluated.
What the study does do is highlight the need for a larger multicentre RCT to address the issue of optimal haemoglobin in a transplant patient and the problems of 'extrapolating' guidelines from one patient group to another.
See Correction of Postkidney Transplant Anemia Reduces Progression of Allograft Nephropathy
Gabriel Choukroun, Nassim Kamar, Bertrand Dussol, Isabelle Etienne, Elisabeth Cassuto-Viguier, Olivier Toupance, François Glowacki, Bruno Moulin, Yvon Lebranchu, Guy Touchard, Maïté Jaureguy, Nicolas Pallet, Yannick Le Meur, Lionel Rostaing, Frank Martinez, and for the CAPRIT study Investigators. J Am Soc Nephrol 2012;23 360-368
Oscar Wilde wrote that " a cynic is someone who knows the price of everything but the value of nothing" ....... Over 100 years later the same might be said of Medicare's attitude to kidney transplantation. In a depressing critique of Medicare's funding of kidney transplantation in the United States, Gill and Tonelli ( see NEJM online first) point out the impact of a lack of a universal funding strategy on clinical outcomes in transplantation. The 10 year graft survival in the US of 43% contrasts with a graft survival of between 55-60% in Canada, Australia and the UK. Whilst these differences maybe due to a multitude of factors it's worth noting that in the US itself the likelihood of graft failure is much higher in the Medicare-insured group than those not insured by Medicare. The risk of graft failure rapidly increases after 3 years in the Medicare group. So what happens at 3 years? At present those insured with Medicare have their funding for immunosuppressants stopped after 3 years unless they are over 65 or deemed to have significant disability. Of course one can't prove that the cost of drugs accounts for the poorer outcomes amongst the Medicare group but the authors point to a survey of US transplant program's in which 68% of centres reported deaths or graft failures related to cost-related non-adherence.
Whilst the clinical reasons to support transplantation are overwhelming the economic reasons are just as robust. The cost of transplantation is initially estimated at $110,000 but this cost falls rapidly after the first year - making transplantation a much more cost effective therapy that dialysis which costs $75,000 per annum with the cumulative cost of dialysis much more than transplantation. However despite the fact that Medicare won't fund immunosuppressants after 3 years ( even though effective generics for Tacrolimus and MMF now widely available) they are able to fund lifelong dialysis. The economic modelling suggests that $200 million could be saved annually by providing universal coverage for lifetime immunosuppressive medication. An amendment to the the Social Security Act before the US Congress proposes to provide universal coverage for transplantation. A previous attempt to provide such coverage in 2009 failed. One can only hope that in the midst of a global economic crisis the policymakers can do their sums and come up with the right answer.
Penny Wise, Pound Foolish? Coverage Limits on Immunosuppression after Kidney Transplantation. John S. Gill, M.D., and Marcello Tonelli, M.D. February 1, 2012 (10.1056/NEJMp1114394) http://www.nejm.org/doi/full/10.1056/NEJMp1114394
The results of the LUNAR study have finally been published in Arthritis and Rheumatism nearly 3 years after being first presented. The study was an RCT of 144 patients looking at the impact of Rituximab (4 doses of 1gm) as add on therapy to steroids and MMF in patients with Class III and IV lupus nephritis. Patients received two doses of methylprednisolone and the steroid taper was from 60mg to around 10 mg at 4 months. The primary endpoint was a composite of CR+PR. At 12 months there was NO significantdifference in renal complete remission (CR) or partial remission (PR). Those in CR/PR were 56.9% in the Rituximab group and 45.8% in the placebo group. CR and PR were rigorously defined by creatinine, proteinuria and haematiuria. Interestingly there was a significant reduction in the C3, C4 and DsDNA (all secondary endpoints) in the Rituximab group and at 72 weeks proteinuria was lower in the Rituximab group. Also in a pre-specified sub-group analysis of race, the primary endpoint was achieved in 70% of black patients receiving Rituximab versus 45% of black patients who received MMF an dsteroids only – this difference wasn’t statistically significant.
This really well designed study seems to suggest there is no role for Rituximab in lupus nephritis in contrast to earlier observational data. Does this mean we should never use Rituximab in lupus nephritis? Well my thoughts are as follows:
- Clearly for most patients with lupus nephritis treated with MMF and steroids, Rituximab will have no added benefit
- We don’t know if there is a role of Rituximab in patients with difficult disease refractory to cyclophosphamide/MMF – as these kinds of patients weren’t in the study. Like most recent studies in LN kidney function was well preserved.
- The study wasn’t really designed to address the role of Rituximab as a steroid minimising/sparing agent as advocated by Imperial College and others. To do this would require a much more aggressive steroid taper than that in the study.
- Finally but importantly as always the choice of endpoint is important in determining whether a study is positive or negative as we have seen in the lipid lowering studies. The study was powered to detect a 25% improvement in renal response rate but this power calculation was heavily skewed towards detecting complete remission rather than partial remission. Thus the authors state that with this sample size the power was less than 70% to detect a 25% improvement in renal response rate where the remissions were predominantly partial rather than complete. Thus whilst the partial remission rate was 31% in the Rituximab group compared to 15% in the placebo group the study was underpowered to detect this difference. Therefore Rituximab maybe having a clinically meaningful effect that this study wasn’t powered to detect.
- Class V LN was excluded and its worth noting that for idiopathic data there is increasing evidence of efficacy of Rituximab
Based on this paper my own practice will continue to be to use MMF and Prednisolone as first line therapy with Cyclophosphamide reserved for those with severe disease. I tend to use Rituximab in i) those who have not responded to MMF/cyclophosphamide ii) those who are unable to take cyclophosphamide ( due to previous exposure or fertility issues) and iii) those who are unable to come off steroids and/or have problems with steroid toxicity.
2. Rituximab is an effective treatment for lupus nephritis and allows a reduction in maintenance steroids.Pepper R, Griffith M, Kirwan C, Levy J, Taube D, Pusey C, Lightstone L, Cairns T. Nephrol Dial Transplant. 2009 Dec;24(12):3717-23. Epub 2009 Jul 17.
- Efficacy and safety of rituximab in patients with active proliferative lupus nephritis: The lupus nephritis assessment with rituximab (LUNAR) study. Rovin BH, Furie R, Latinis K, Looney RJ, Fervenza FC, Sanchez-Guerrero J, Maciuca R, Zhang D, Garg JP, Brunetta P, Appel G; for the LUNAR Investigator Group. Arthritis Rheum. 2012 Jan 9. doi: 10.1002/art.34359.
Anyone who has been to a GKA masterclass will know that a key learning message from Professor El Nahas' session on 'how to read a paper' is to assess whether the findings are relevant to your healthcare setting. I was thinking about this recently when a large RCT was recently published in NEJM evaluating whether low molecular weight heparin reduces mortality in acutely ill medical patients- this was of interest as in the UK we are aggressively audited to make sure all acute admissions get prophylactic heparin. The outcome of this trial of more than 8000 patients was negative but what interested me was where the trial was conducted - China, India, Korea, Malaysia, Mexico, the Philippines, and Tunisia. One question that arises is how applicable this data would be to say a Uk setting? We know that in the UK there are huge variations in outcome for the same diseases in different hospitals- despite that fact that we have only one major supplier of healthcare - the NHS. So I imagine both care and patient characteristics maybe very different ( not better or worse just different) in the UK than say Tunisia. Similarly in two trials of belimumab in SLE the BLISS 52 trial reported relatively positive outcomes whilst BLISS 76 seemed to suggest that belimumab had a more marginal effect. Some have suggested that this was because BLISS 76 was conducted predominantly in the USA and Europe whilst Bliss 52 was conducted predominantly in Asia and Eastern Europe - the implication being that either the disease itself or standard care is different between say the USA and Asia My own thoughts are that all we can do is look at data and critically evaluate the relevance to the healthcare setting we practice in. Of course it also suggests that we all have a responsibility to try if possible to get involved in clinical research conducted in line with good clinical practice to ensure evidence-based decision making continues to develop.
See : http://www.nejm.org/doi/full/10.1056/NEJMoa1111288?query=TOC
With the epidemic of type 2 diabetes, obesity and an ageing population much of modern day nephrology focusses primarily on cardiovascular risk management - i.e. blood pressure, lipid and glycaemic control. One question I've always asked myself is whether the traditional model of a patient visiting a doctor in a clinic at two to three monthly intervals is clinically useful or cost effective. Many of the lifestyle interventions such as diet, exercise, stopping smoking, salt restriction etc dont actually require specialist clinical input whilst for many patients algorithms for BP, lipid or glycaemic control are relatively simple and well established through national guidelines. Indeed in the UK, some specialist nurses appropriately trained in prescribing having taken on an important role managing diabetes and hypertension in primary care. Furthermore it is clear that in middle and low income countries ( where there is a real epidemic of non-communicable diseases (NCDs)) such a doctor-centric model of care is simply not financially viable.
In this weeks Lancet there is an interesting analysis of the Iranian approach to NCDs with an analysis of their approach to rural primary health care. Their model known as the Behvarz system uses trained community health care workers to serve rural populations. These workers are trained for 2 years and receive a salary that is 1/6th that of a doctors salary. Between 1996 and 2002 as part of a national action plan for the prevention and control of diabetes, Behvarz workers were trained to identify those at high risk of diabetes, provide lifestyle advise and follow up those diagnosed with diabetes to monitor for complications and promote adherence and identify patients who are running into problems to doctors . Data from the NCD surveillance survey in 2005 ( a large health examination survey of nearly 90,000 individuals aged between 15-64) was combined with socio-economic data and data on the density of Behvarz workers to study the impact of the scheme. Treatment reduced plasma glucose by 1.34 mmol/l in rural (i.e. Behvarz areas) compared to 0.21 mmol/l in urban areas. With respect to blood pressure the results were slightly better in urban areas though its important to appreciate that blood pressure management had not been incorporated into the Behvarz system.
Whilst the study is not a RCT of two different models of care this large, high quality dataset makes interesting reading. The better control of diabetes in rural areas did seem to be related to the use of Behvarz workers who perhaps played a key role in ensuring adherence to therapeutic plans. If healthcare systems around the world are to rise to the challenge of NCDs then we will need to develop alternative, cost-effective models of care - and we will need to challenge whether what doctors do can be delivered (or complemented) by other healthcare workers in a more cost-effective way.
see Farzadfar and colleagues http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(11)61349-4/abstract
So after analysing the effect of obesity and phosphorus yet another post-hoc analysis from the REIN study is about to be published in JASN ! This analysis looked at the effect of dietary salt intake (as measured by 24 hour urinary sodium and creatinine) suggests that those with high salt intake were more likely to progress to ESRD and that this was mediated by attenuating the anti-proteinuric effects of ACE-I. Among those with low, medium, and high sodium intakes, the incidence of ESRD was 6.1 %, 7.9 and 18.2 per 100 patient-years, respectively (P<0.001). Somewhat counter-intuitively there was no difference in blood pressure and the association between salt intake and progression was lost when taking into account changes in proteinuria during follow up - leading the authors to conclude that the anti-proteinuric effects of ACE-i were mitigated by salt intake.The authors suggest that expansion of the sodium pool leads to glomerular hyperfiltration and activation of the renal RAS ithereby mitigating the anti-proteinuric effects of ACEi.
It has been established for many years that diuretics and salt restriction both augment the anti-hypertensive effects of ACE-i and so it is surprising to see no difference in BP between the hogh and low sodium groups. As said before I am never comfortable with the use of post-hoc analysis to try to elucudate the mechanism of action of a drug or a treatment recommendation. However the article reinforces the need salt restriction in the CKD population and the importance of factoring salt intake into any trials of progression. Whilst salt restriction is clearly important I'm not sure I share the authors belief that salt restriction can be achieved easily by 'small inconveniences of minimal dietary restrictions" - my personal experiene is that all dietary interventions are difficult to achieve and require considerable dietic resource.