- Email: [email protected]
Chronic kidney disease in Taiwan
Correspondence
Cutaneous blood-filled vesicles on idraparinux In the Amadeus study,1 patients in atrial fibrillation who had an increased risk of thromboembolic stroke were randomly assigned either a vitamin K antagonist or a weekly subcutaneous injection of idraparinux—a synthetic pentasaccharide that inhibits activated factor X. The study was stopped early because of excessive bleeding in patients assigned idraparinux. We report unusual skin lesions that occurred in 15 of 56 participants assigned open-label idraparinux compared with none of 59 patients assigned warfarin at our study site. After noting skin lesions in two study patients, we did detailed skin examinations on all study participants at each follow-up visit. We saw raised, blood-filled vesicles 0·5–2·0 cm in diameter that were remote from the subcutaneous injection sites in 15 patients (figure). Most patients had two to eight lesions, usually on the arms and legs, which appeared on average 3 months (range 2–8) after starting idraparinux. From this time new lesions continued to appear, but the severity of the lesions did not increase despite ongoing medication. On first appearance the lesions would be bright red, suggesting fresh blood. They would then darken before gradually resolving over 2 weeks. The lesions were not painful or itchy, and occurred spontaneously with no trauma. The skin lesions continued after onset during a median follow-
Figure: Blood-filled vesicle measuring 1 cm in diameter on the leg of a patient treated with idraparinux
www.thelancet.com Vol 372 December 6, 2008
up of 5 months (range 1–10). In two patients the lesions disappeared over the course of 2 weeks after temporarily stopping idraparinux, but returned within 1 month on restarting idraparinux. In all patients, new lesions stopped appearing and gradually resolved 1 week after the last injection of idraparinux at study end with no subsequent recurrence.2 These skin lesions did not meet the study definition of a clinically significant bleed, which included a subcutaneous haematoma of more than 25 cm² or more than 100 cm² after trauma. Patients with vesicular skin lesions were no more likely to have a clinically significant bleed than were idraparinux patients with no skin lesions (four of 15 vs 15 of 43), although both groups were more likely to bleed than were patients assigned warfarin (seven of 59). We found no evidence of other systemic abnormalities in patients with skin lesions, and the platelet count remained in the normal range for all patients. These skin lesions are likely to be a specific adverse effect of idraparinux. Raised, blood-filled vesicles with no history of trauma are not typical of subcutaneous bleeding seen with vitamin K antagonists, antiplatelet drugs, and other antithrombotic treatments. Raised purpuric lesions can occur with hypersensitivity vasculitis to drugs,3 but the absence of local itching or burning, systemic symptoms, and lack of progressive severity make this diagnosis less likely. Skin biopsy was not done in our patients, and the pathophysiology and clinical significance of these lesions is currently uncertain. These skin lesions have not been reported in previous studies of idraparinux2,4,5 or fondaparinux (another pentasaccharide inhibitor of factor Xa3). Future studies of idraparinux should include documentation of the occurrence of skin lesions, and additional investigations to determine their pathology. JB and RAHS were investigators in the Amadeus study and have no conflicts of interest to declare.
*J Benatar, R A H Stewart [email protected] Green Lane Cardiovascular Service, Auckland City Hospital, Private Bag 92189, Auckland, New Zealand 1
2
3
4
5
The Amadeus Investigators. Comparison of idraparinux with vitamin K antagonists for prevention of thromboembolism in patients with atrial fibrillation: a randomised, open-label, non-inferiority trial. Lancet 2008; 371: 315–21. Buller HR, Cohen AT, Davidson B, et al. Idraparinux versus standard therapy for venous thromboembolic disease. N Engl J Med 2007; 357: 1094–104. ten Holder SM, Joy MS, Falk RJ. Cutaneous and systemic manifestations of drug-induced vasculitis. Ann Pharmacother 2002; 36: 130–47. Bauer KA, Eriksson BI, Lassen MR, Turpie AG. Fondaparinux compared with enoxaparin for the prevention of venous thromboembolism after elective major knee surgery. N Engl J Med 2001; 345: 1305–10. Buller HR, Cohen AT, Davidson B, et al. Extended prophylaxis of venous thromboembolism with idraparinux. N Engl J Med 2007; 357: 1105–12.
Chronic kidney disease in Taiwan Chi Peng Wen and colleagues (June 28, p 2173)1 are to be commended on their massive undertaking to characterise kidney disease and its potential effect on all-cause mortality in Taiwanese adults. However, we are concerned that the analysis is flawed and gives a misleading message. The study participants were not representative of the general population but were fee-paying, with incentives given to members of large families. Chronic kidney disease (CKD) is known to prevail within families. Additionally, the National Kidney Foundation Kidney Disease Outcomes Quality Initiative2 overestimates the prevalence of CKD owing to its use of an absolute threshold of estimated glomerular filtration rate (eGFR) for defining stage 3 CKD (30–59 mL/min/1·73 m²) without any adjustment for the effects of normal ageing on eGFR.3 We also question whether the excretion of small amounts of protein in the urine should be used to define CKD in the absence of other findings. If one assumes that the isolated finding of minimal proteinuria is not indicative of CKD and further that at least half stage
Submissions should be made via our electronic submission system at http://ees.elsevier.com/ thelancet/
1949
Correspondence
3a CKD is a result of normal ageing, the estimated prevalence of significant CKD falls from 11·9% to about 3·7%. Wen and colleagues’ cohort contained many elderly individuals (16% of the entire cohort were older than 55 years at recruitment and more than 50% of those with stage 3 CKD were older than 60 years). The hazard ratios for all-cause mortality for patients categorised as having stage 3a disease on the basis of eGFR (45–59 mL/min/1·73 m²) and who did not have abnormal proteinuria were no different from those for patients with “normal” eGFR (>90 mL/min/1·73 m²). These data suggest that the effect of a low eGFR in the absence of abnormal proteinuria is not seen until the eGFR falls below the normal levels expected after adjustment for age and gender effects (about 45 mL/min/1·73 m²). The true prevalence of CKD has been greatly overestimated in this study and the implications of the existence of a “global epidemic” of CKD could be incorrect and misleading. We declare that we have no conflict of interest.
*Richard J Glassock, Meguid El Nahas, Christopher G Winearls [email protected] David Geffen School of Medicine at UCLA, Los Angeles, CA 92677, USA (RJG); Sheffield Kidney Institute; Sheffield, UK (MEN); and Oxford Kidney Unit, Oxford, UK (CGW) 1
2
3
Wen CP, Cheng TYD, Tsai MK, et al. All-cause mortality attributable to chronic kidney disease: a prospective cohort study based on 462 293 adults in Taiwan. Lancet 2008; 371: 2173–82. National Kidney Foundation. Kidney Disease Outcomes Quality Initiative. Am J Kidney Dis 2002; (suppl 1): s17–31. Wetzels JFM, Willems HL, den Heijer M. Age- and gender-specific reference values of estimated glomerular filtration rate in a Caucasian population: results of the Nijmegen Biomedical Study. Kidney Int 2008; 73: 656–58.
(GFR). Even if such non-calibration has little importance for mortality linked to CKD, it could have serious consequences on prevalence data.2 These prevalence data are thus not easy to compare with those of the US population, for which the newly expressed Modification of Diet in Renal Disease (MDRD) study equation3 was used. This equation is applied to standardised creatinine with a factor of 175; Wen and colleagues used the version of the equation with a factor of 186. Moreover, the relation between creatinine and GFR varies with ethnic origin. For African-American3 and Japanese4 populations, correction factors of 1·21 and 0·763, respectively, must be applied. Such a correction was not done by Wen and colleagues and should be discussed. Lastly, Wen and colleagues show that nearly one in three patients older than 65 years presented with stage 3 CKD. The term “disease” in this age group is debatable because the normal GFR in older populations is not well defined; a GFR of less than 60 mL/min/1·73 m² could be regarded as physiological in a healthy older person.5 We declare that we have no conflict of interest.
*Pierre Delanaye, Etienne Cavalier, Jean-Marie Krzesinski [email protected] Department of Dialysis-Nephrology (PD, J-MK) and Department of Clinical Chemistry (EC), University of Liège, CHU Sart Tilman, 4000 Liège, Belgium 1
2
3
In Chi Peng Wen and colleagues’ paper on the mortality attributable to chronic kidney disease (CKD) in Taiwan,1 the prevalence of CKD must be interpreted with caution. Wen and colleagues use a noncalibrated creatinine concentration to estimate glomerular filtration rate 1950
4
5
Wen CP, Cheng TYD, Tsai MK, et al. All-cause mortality attributable to chronic kidney disease: a prospective cohort study based on 462 293 adults in Taiwan. Lancet 2008; 371: 2173–82. Coresh J, Eknoyan G, Levey AS. Estimating the prevalence of low glomerular filtration rate requires attention to the creatinine assay calibration. J Am Soc Nephrol 2002; 13: 2811–12. Levey AS, Coresh J, Greene T, et al. Using standardized serum creatinine values in the Modification of Diet in Renal Disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006; 145: 247–54. Imai E, Horio M, Nitta K, et al. Modification of the Modification of Diet in Renal Disease (MDRD) Study equation for Japan. Am J Kidney Dis 2007; 50: 927–37. Glassock RJ, Winearls C. An epidemic of chronic kidney disease: fact or fiction? Nephrol Dial Transplant 2008; 23: 1117–21.
Authors’ reply Richard Glassock and colleagues ask whether reduced glomerular filtration rate (GFR) is a normal ageing process and whether our cohort could be skewed by including a larger number of older people in it. Older people were not over-represented in our cohort: 21·7% were aged 55 years or older versus 24·9% in the general population (corresponding figures for age 65 years or older 8·0% vs 13·4%). As shown in the table, those with stage 3a disease (GFR 45–59 mL/min/1·73 m²) had significantly increased mortality, including those with negative urine protein. The increased risk of allcause mortality signifies that these individuals have a shortened lifespan and thus the reduced GFR could hardly be considered a sign of normal ageing. The age debate reminded us of the one on whether cut-points for hypertension should be age-adjusted decades ago. The practice of associating older age with higher blood pressure was abandoned once mortality outcome data became available. Glassock and colleagues are also concerned about the accuracy of CKD prevalence based on only one test. We found the adverse outcome of significantly increased mortality in those with abnormal findings in a single investigation, either urine test or estimated GFR, so this group with increased risks deserves our clinical attention. Even in the face of this, we have downwardly adjusted the prevalence, a process also used in arriving at the US prevalence,1 so the results were more on the conservative side than otherwise. Clinicians tend to favour repeated testing—eg, advocating two of three tests2—but this three-test scenario is cumbersome and defeats the purpose of screening seemingly healthy individuals—ie, the majority of patients with CKD. The two-test scenario also adds to the confusion as to why a second test is more valid than the first, when they are inconsistent. In answer to Pierre Delanaye and colleagues’ concern, we carefully www.thelancet.com Vol 372 December 6, 2008
Cutaneous blood-filled vesicles on idraparinux In the Amadeus study,1 patients in atrial fibrillation who had an increased risk of thromboembolic stroke were randomly assigned either a vitamin K antagonist or a weekly subcutaneous injection of idraparinux—a synthetic pentasaccharide that inhibits activated factor X. The study was stopped early because of excessive bleeding in patients assigned idraparinux. We report unusual skin lesions that occurred in 15 of 56 participants assigned open-label idraparinux compared with none of 59 patients assigned warfarin at our study site. After noting skin lesions in two study patients, we did detailed skin examinations on all study participants at each follow-up visit. We saw raised, blood-filled vesicles 0·5–2·0 cm in diameter that were remote from the subcutaneous injection sites in 15 patients (figure). Most patients had two to eight lesions, usually on the arms and legs, which appeared on average 3 months (range 2–8) after starting idraparinux. From this time new lesions continued to appear, but the severity of the lesions did not increase despite ongoing medication. On first appearance the lesions would be bright red, suggesting fresh blood. They would then darken before gradually resolving over 2 weeks. The lesions were not painful or itchy, and occurred spontaneously with no trauma. The skin lesions continued after onset during a median follow-
Figure: Blood-filled vesicle measuring 1 cm in diameter on the leg of a patient treated with idraparinux
www.thelancet.com Vol 372 December 6, 2008
up of 5 months (range 1–10). In two patients the lesions disappeared over the course of 2 weeks after temporarily stopping idraparinux, but returned within 1 month on restarting idraparinux. In all patients, new lesions stopped appearing and gradually resolved 1 week after the last injection of idraparinux at study end with no subsequent recurrence.2 These skin lesions did not meet the study definition of a clinically significant bleed, which included a subcutaneous haematoma of more than 25 cm² or more than 100 cm² after trauma. Patients with vesicular skin lesions were no more likely to have a clinically significant bleed than were idraparinux patients with no skin lesions (four of 15 vs 15 of 43), although both groups were more likely to bleed than were patients assigned warfarin (seven of 59). We found no evidence of other systemic abnormalities in patients with skin lesions, and the platelet count remained in the normal range for all patients. These skin lesions are likely to be a specific adverse effect of idraparinux. Raised, blood-filled vesicles with no history of trauma are not typical of subcutaneous bleeding seen with vitamin K antagonists, antiplatelet drugs, and other antithrombotic treatments. Raised purpuric lesions can occur with hypersensitivity vasculitis to drugs,3 but the absence of local itching or burning, systemic symptoms, and lack of progressive severity make this diagnosis less likely. Skin biopsy was not done in our patients, and the pathophysiology and clinical significance of these lesions is currently uncertain. These skin lesions have not been reported in previous studies of idraparinux2,4,5 or fondaparinux (another pentasaccharide inhibitor of factor Xa3). Future studies of idraparinux should include documentation of the occurrence of skin lesions, and additional investigations to determine their pathology. JB and RAHS were investigators in the Amadeus study and have no conflicts of interest to declare.
*J Benatar, R A H Stewart [email protected] Green Lane Cardiovascular Service, Auckland City Hospital, Private Bag 92189, Auckland, New Zealand 1
2
3
4
5
The Amadeus Investigators. Comparison of idraparinux with vitamin K antagonists for prevention of thromboembolism in patients with atrial fibrillation: a randomised, open-label, non-inferiority trial. Lancet 2008; 371: 315–21. Buller HR, Cohen AT, Davidson B, et al. Idraparinux versus standard therapy for venous thromboembolic disease. N Engl J Med 2007; 357: 1094–104. ten Holder SM, Joy MS, Falk RJ. Cutaneous and systemic manifestations of drug-induced vasculitis. Ann Pharmacother 2002; 36: 130–47. Bauer KA, Eriksson BI, Lassen MR, Turpie AG. Fondaparinux compared with enoxaparin for the prevention of venous thromboembolism after elective major knee surgery. N Engl J Med 2001; 345: 1305–10. Buller HR, Cohen AT, Davidson B, et al. Extended prophylaxis of venous thromboembolism with idraparinux. N Engl J Med 2007; 357: 1105–12.
Chronic kidney disease in Taiwan Chi Peng Wen and colleagues (June 28, p 2173)1 are to be commended on their massive undertaking to characterise kidney disease and its potential effect on all-cause mortality in Taiwanese adults. However, we are concerned that the analysis is flawed and gives a misleading message. The study participants were not representative of the general population but were fee-paying, with incentives given to members of large families. Chronic kidney disease (CKD) is known to prevail within families. Additionally, the National Kidney Foundation Kidney Disease Outcomes Quality Initiative2 overestimates the prevalence of CKD owing to its use of an absolute threshold of estimated glomerular filtration rate (eGFR) for defining stage 3 CKD (30–59 mL/min/1·73 m²) without any adjustment for the effects of normal ageing on eGFR.3 We also question whether the excretion of small amounts of protein in the urine should be used to define CKD in the absence of other findings. If one assumes that the isolated finding of minimal proteinuria is not indicative of CKD and further that at least half stage
Submissions should be made via our electronic submission system at http://ees.elsevier.com/ thelancet/
1949
Correspondence
3a CKD is a result of normal ageing, the estimated prevalence of significant CKD falls from 11·9% to about 3·7%. Wen and colleagues’ cohort contained many elderly individuals (16% of the entire cohort were older than 55 years at recruitment and more than 50% of those with stage 3 CKD were older than 60 years). The hazard ratios for all-cause mortality for patients categorised as having stage 3a disease on the basis of eGFR (45–59 mL/min/1·73 m²) and who did not have abnormal proteinuria were no different from those for patients with “normal” eGFR (>90 mL/min/1·73 m²). These data suggest that the effect of a low eGFR in the absence of abnormal proteinuria is not seen until the eGFR falls below the normal levels expected after adjustment for age and gender effects (about 45 mL/min/1·73 m²). The true prevalence of CKD has been greatly overestimated in this study and the implications of the existence of a “global epidemic” of CKD could be incorrect and misleading. We declare that we have no conflict of interest.
*Richard J Glassock, Meguid El Nahas, Christopher G Winearls [email protected] David Geffen School of Medicine at UCLA, Los Angeles, CA 92677, USA (RJG); Sheffield Kidney Institute; Sheffield, UK (MEN); and Oxford Kidney Unit, Oxford, UK (CGW) 1
2
3
Wen CP, Cheng TYD, Tsai MK, et al. All-cause mortality attributable to chronic kidney disease: a prospective cohort study based on 462 293 adults in Taiwan. Lancet 2008; 371: 2173–82. National Kidney Foundation. Kidney Disease Outcomes Quality Initiative. Am J Kidney Dis 2002; (suppl 1): s17–31. Wetzels JFM, Willems HL, den Heijer M. Age- and gender-specific reference values of estimated glomerular filtration rate in a Caucasian population: results of the Nijmegen Biomedical Study. Kidney Int 2008; 73: 656–58.
(GFR). Even if such non-calibration has little importance for mortality linked to CKD, it could have serious consequences on prevalence data.2 These prevalence data are thus not easy to compare with those of the US population, for which the newly expressed Modification of Diet in Renal Disease (MDRD) study equation3 was used. This equation is applied to standardised creatinine with a factor of 175; Wen and colleagues used the version of the equation with a factor of 186. Moreover, the relation between creatinine and GFR varies with ethnic origin. For African-American3 and Japanese4 populations, correction factors of 1·21 and 0·763, respectively, must be applied. Such a correction was not done by Wen and colleagues and should be discussed. Lastly, Wen and colleagues show that nearly one in three patients older than 65 years presented with stage 3 CKD. The term “disease” in this age group is debatable because the normal GFR in older populations is not well defined; a GFR of less than 60 mL/min/1·73 m² could be regarded as physiological in a healthy older person.5 We declare that we have no conflict of interest.
*Pierre Delanaye, Etienne Cavalier, Jean-Marie Krzesinski [email protected] Department of Dialysis-Nephrology (PD, J-MK) and Department of Clinical Chemistry (EC), University of Liège, CHU Sart Tilman, 4000 Liège, Belgium 1
2
3
In Chi Peng Wen and colleagues’ paper on the mortality attributable to chronic kidney disease (CKD) in Taiwan,1 the prevalence of CKD must be interpreted with caution. Wen and colleagues use a noncalibrated creatinine concentration to estimate glomerular filtration rate 1950
4
5
Wen CP, Cheng TYD, Tsai MK, et al. All-cause mortality attributable to chronic kidney disease: a prospective cohort study based on 462 293 adults in Taiwan. Lancet 2008; 371: 2173–82. Coresh J, Eknoyan G, Levey AS. Estimating the prevalence of low glomerular filtration rate requires attention to the creatinine assay calibration. J Am Soc Nephrol 2002; 13: 2811–12. Levey AS, Coresh J, Greene T, et al. Using standardized serum creatinine values in the Modification of Diet in Renal Disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006; 145: 247–54. Imai E, Horio M, Nitta K, et al. Modification of the Modification of Diet in Renal Disease (MDRD) Study equation for Japan. Am J Kidney Dis 2007; 50: 927–37. Glassock RJ, Winearls C. An epidemic of chronic kidney disease: fact or fiction? Nephrol Dial Transplant 2008; 23: 1117–21.
Authors’ reply Richard Glassock and colleagues ask whether reduced glomerular filtration rate (GFR) is a normal ageing process and whether our cohort could be skewed by including a larger number of older people in it. Older people were not over-represented in our cohort: 21·7% were aged 55 years or older versus 24·9% in the general population (corresponding figures for age 65 years or older 8·0% vs 13·4%). As shown in the table, those with stage 3a disease (GFR 45–59 mL/min/1·73 m²) had significantly increased mortality, including those with negative urine protein. The increased risk of allcause mortality signifies that these individuals have a shortened lifespan and thus the reduced GFR could hardly be considered a sign of normal ageing. The age debate reminded us of the one on whether cut-points for hypertension should be age-adjusted decades ago. The practice of associating older age with higher blood pressure was abandoned once mortality outcome data became available. Glassock and colleagues are also concerned about the accuracy of CKD prevalence based on only one test. We found the adverse outcome of significantly increased mortality in those with abnormal findings in a single investigation, either urine test or estimated GFR, so this group with increased risks deserves our clinical attention. Even in the face of this, we have downwardly adjusted the prevalence, a process also used in arriving at the US prevalence,1 so the results were more on the conservative side than otherwise. Clinicians tend to favour repeated testing—eg, advocating two of three tests2—but this three-test scenario is cumbersome and defeats the purpose of screening seemingly healthy individuals—ie, the majority of patients with CKD. The two-test scenario also adds to the confusion as to why a second test is more valid than the first, when they are inconsistent. In answer to Pierre Delanaye and colleagues’ concern, we carefully www.thelancet.com Vol 372 December 6, 2008