- Email: [email protected]
Limitations of BOLD-MRI for assessment of hypoxia in chronically diseased human kidneys
letter to the editor
http://www.kidney-international.org & 2012 International Society of Nephrology
Is there no future for renal BOLD-MRI? To the Editor: We read with interest the recent article by Michaely et al.1 in which it was stated that renal blood oxygen level-dependent magnetic resonance imaging (BOLD-MRI) does not reflect renal function in chronic kidney disease (CKD), a conclusion that is in direct contrast to our recent results.2 We strongly support the statement3 from the same issue that renal tissue oxygenation appears to be dependent on CKD severity as well as on the etiology of the underlying kidney disease. The causes of chronic renal hypoxia are remarkably multifactorial,4 and while we have verified a relationship between BOLD-MRI T2* values and estimated glomerular filtration rates in nondiabetic nephropathy, no association has been shown in cases of diabetic nephropathy.2 Thus, etiological diversity in CKD might have a far greater, and more pernicious, influence on BOLD-MRI results than thought by Michaely et al. As the kidney is surrounded by adipose tissue, removal of the out-of-phase subtraction effect caused by this tissue is vital to minimizing fluctuations in T2* values. We, therefore, used in-phase echo time (TE) of the longest possible duration to limit the T2* errors and provide more accurate data, while balancing the signal-to-noise ratio. Michaely et al. used an unusually short TE of just 40 ms. We, and other similarly focused groups, understand the limitations of this novel modality and hope that the article does not discourage the application of BOLD-MRI to CKD assessment. 1. Michaely HJ, Metzger L, Haneder S et al. Renal BOLD-MRI does not reflect renal function in chronic kidney disease. Kidney Int 2012; 81: 684–689. 2. Inoue T, Kozawa E, Okada H et al. Noninvasive evaluation of kidney hypoxia and fibrosis using magnetic resonance imaging. J Am Soc Nephrol 2011; 22: 1429–1434. 3. Neugarten J. Renal BOLD-MRI and assessment for renal hypoxia. Kidney Int 2012; 81: 613–614. 4. Mimura I, Nangaku M. The suffocating kidney: tubulointerstitial hypoxia in end-stage renal disease. Nat Rev Nephrol 2010; 6: 667–678.
Tsutomu Inoue1, Eito Kozawa2, Hirokazu Okada1 and Hiromichi Suzuki1 1 Department of Nephrology, Faculty of Medicine, Saitama Medical University, Saitama, Japan and 2Department of Diagnostic Radiology, Saitama Medical University International Medical Center, Saitama, Japan Correspondence: Tsutomu Inoue, Department of Nephrology, Faculty of Medicine, Saitama Medical University, 38 Morohongo, Moroyama-machi, Iruma-gun, Saitama, Japan. E-mail: [email protected]
(CKD) with different degrees of severity of renal failure, in which blood oxygen level–dependent magnetic resonance imaging (BOLD-MRI) was used as a non-invasive technique for monitoring the degree of renal oxygenation. This study addressed the contention that intrarenal chronic hypoxia underlies the progressive nature of most forms of CKD.2–4 The authors were unable to discriminate between patients with different stages of CKD using this method. We would argue that BOLD-MRI is not the appropriate method for assessing oxygenation in CKD given the specific evolution of the pathology in this setting. BOLD-MRI measures vascular oxygen saturation with the signal being indirectly related to deoxyhemoglobin concentration.5,6 To compare different stages of disease for the degree of hypoxia, the volume of the vessels and the number of red blood cells must be the same in the voxels being compared, to make the comparison valid. The patchy involvement of the chronically diseased kidney, in which microvascular obliteration occurs in the tubule-interstitial compartment, means that the expression of a single datum point to reflect the entire cortex or medulla (as portrayed in the paper by Michaely et al.) would fail to capture this heterogeneity in the early and mid-stages of the disease. Expressing the variability of the BOLD signal intensity or R2* of pixels contained within the measured region of interest for each patient may be a more appropriate way to detect differences. Equally important, in more advanced disease, both microvasculature loss and a diminished red blood cell mass occur as CKD progresses. Hence, the source of the signal for deoxyhemoglobin concentration in a volume being measured would be diluted as the disease progresses, thereby obscuring functionally relevant local hypoxic regions and making comparisons with less advanced disease invalid. We concur with the author of the accompanying editorial that ‘limitations in the technique and execution’ lead to the conclusion that ‘the hypoxia hypothesis has not been refuted’.7 The study is important, however, for raising the need for more refined use of BOLD-MRI or for alternative methods for determining intrarenal hypoxia in CKD. 1.
2.
Kidney International (2012) 82, 934; doi:10.1038/ki.2012.282 3. 4.
Limitations of BOLD-MRI for assessment of hypoxia in chronically diseased human kidneys To the Editor: The article by Michaely et al.1 describes a sizeable population of patients with chronic kidney disease 934
5.
6.
7.
Michaely HJ, Metzger L, Haneder S et al. Renal BOLD-MRI does not reflect renal function in chronic kidney disease. Kidney Int 2012; 81: 684–689. Fine LG, Norman JT. Chronic hypoxia as a mechanism of progression of chronic kidney diseases: from hypothesis to novel therapeutics. Kidney Int 2008; 74: 867–872. Nangaku M. Chronic hypoxia and tubulointerstitial injury: a final common pathway to end-stage renal failure. J Am Soc Nephrol 2006; 17: 17–25. Long DA, Norman JT, Fine LG. Restoring the renal microvasculature to treat chronic kidney disease. Nat Rev Nephrol 2012; 8: 244–250. Mangalathu Arumana J, Li D, Dharmakumar R. Deriving blood-oxygenlevel-dependent contrast in MRI: an evaluation of T2*-weighted, T2-prepared and SSFP methods at 1.5 T and 3.0 T. Magn Reson Med 2008; 27: 1037–1045. Dharmakumar R, Qi X, Hong J et al. Detecting microcirculatory changes in blood oxygen state with steady-state free precession imaging. Magn Reson Med 2006; 55: 1372–1380. Neugarten J. Renal BOLD-MRI and assessment for renal hypoxia. Kidney Int 2012; 81: 613–614. Kidney International (2012) 82, 934–937
letter to the editor
Leon G. Fine1 and Rohan Dharmakumar1
5.
1
Department of Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California, USA Correspondence: Leon G. Fine, Department of Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California 90048, USA. E-mail: [email protected] Kidney International (2012) 82, 934–935; doi:10.1038/ki.2012.283
Gloviczki ML, Glockner J, Gomez SI et al. Comparison of 1.5 and 3 T BOLD MR to study oxygenation of kidney cortex and medulla in human renovascular disease. Invest Radiol 2009; 44: 566–557.
Henrik J.M. Michaely1 1
Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Mannheim, Germany Correspondence: Henrik J.M. Michaely, Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Mannheim, BA-Wu¨, Germany. E-mail: [email protected] Kidney International (2012) 82, 935; doi:10.1038/ki.2012.284
The Author Replies: In their letters to the editor, two groups, from Los Angeles1 and Saitama, Japan,2 respectively, raised some concerns regarding the methodology and the results of our previously published study on renal blood oxygen level– dependent magnetic resonance imaging (BOLD-MRI) in a non-selected patient group. The current literature on renal BOLD-MRI and its applications presents heterogeneous results regarding the correlation between R2* and the estimated glomerular filtration rate.3,4 These different results are most likely due to a lack of understanding of the complex interaction between the different types of renal diseases, their specific pathophysiology, and the complex magnetic resonance (MR) signal generation. The choice of shorter echo times in our study is surely not a relevant factor explaining the differences between our findings and the findings of other research groups. In contrast, the use of 3T MR mandates the use of shorter echo times to yield reliable image quality even with the longest echo time used.5 In addition, similar echo times as in our study have also been used by the group of Prasad et al., who pioneered the field of renal BOLD-MRI. Much more interesting than these technical details are the implications of our study that have been addressed by the authors of these letters. We concur that our study does not imply that BOLD-MRI of the kidneys should not be done. The main conclusion of our study was that ‘broad application of renal BOLD MRI seems not suitable’.3 The use of BOLD imaging in dedicated patient groups should not be discouraged. In contrast, further studies in patient groups with similar pathophysiological changes or intraindividual studies are warranted. Under these premises the influence of altered microvascular density or of the red blood cell count and altered medications could be better controlled. Novel approaches for image analysis of the parametric R2* maps, such as the suggested regionof-interest–based histogram plot, might further provide better insights into pathology. 1. Fine L, Dharmakumar R. Limitations of BOLD-MRI for assessment of hypoxia. Kidney Int 2012; 82: 934–955. 2. Inoue T, Kozawa E, Okada H et al. Is there no future for renal BOLD MRI? Kidney Int 2012; 82: 934. 3. Michaely HJ, Metzger L, Haneder S et al. Renal BOLD-MRI does not reflect renal function in chronic kidney disease. Kidney Int 2012; 81: 684–689. 4. Inoue T, Kozawa E, Okada H et al. Noninvasive evaluation of kidney hypoxia and fibrosis using magnetic resonance imaging. J Am Soc Nephrol 2011; 22: 1429–1434. Kidney International (2012) 82, 934–937
Drug use and nephrotoxicity in the intensive care unit: are we under-dosing antimicrobials in patients with acute kidney injury with the need of extended dialysis? To the Editor: An important aspect that we would like to raise, in addition to what Dr Perazella described as drug dose adjustment1 in a recent review in this journal, is the use of antibiotics in extended dialysis (ED). With ED, patients with acute kidney injury have a potentially supraphysiologic drug clearance during ED procedures, and this may be detrimental in patients for whom bactericidal concentrations of antibiotics are needed. Therefore, when the antibiotic is given in relation to ED may be as important as the dosing. This premise has been demonstrated by Lorenzen et al.,2 who reported that ampicillin/ sulbactam given too early or late in relation to ED could result in supra- or subtherapeutic concentrations. Also considered should be the pharmacodynamic properties of the antimicrobial in question. Antimicrobials require specific attention to their pharmacodynamic activity, as drugs that are time-dependent killers may require a different timing and dosing compared with those that kill in a concentration-dependent manner.3 Taking these considerations into account may be difficult enough, but complexity is added by the intensive care unit (ICU) environment, to which access is limited, and procedures may interfere with drug and dialysis scheduling. In addition, there is paucity in dosing information of several antibiotics commonly used in the ICU in patients with ED. ED pharmacokinetics has been conducted in very few drugs.4 We agree with Dr Perazella that dosing of medications in critically ill patients with acute renal replacement therapy is challenging, and that practitioners should seek best evidence and adjust their recommendation of dialysis parameters and timing to optimize the dosing of antimicrobials. 1. 2.
Perazella MA. Drug use and nephrotoxicity in the intensive care unit. Kidney Int 2012; 81: 1172–1178. Lorenzen JM, Broll M, Kaever V et al. Pharmacokinetics of ampicillin/ sulbactam in critically ill patients with acute kidney injury undergoing to extended dialysis. Clin J Am Soc Nephrol 2012; 7: 385–389.
935
http://www.kidney-international.org & 2012 International Society of Nephrology
Is there no future for renal BOLD-MRI? To the Editor: We read with interest the recent article by Michaely et al.1 in which it was stated that renal blood oxygen level-dependent magnetic resonance imaging (BOLD-MRI) does not reflect renal function in chronic kidney disease (CKD), a conclusion that is in direct contrast to our recent results.2 We strongly support the statement3 from the same issue that renal tissue oxygenation appears to be dependent on CKD severity as well as on the etiology of the underlying kidney disease. The causes of chronic renal hypoxia are remarkably multifactorial,4 and while we have verified a relationship between BOLD-MRI T2* values and estimated glomerular filtration rates in nondiabetic nephropathy, no association has been shown in cases of diabetic nephropathy.2 Thus, etiological diversity in CKD might have a far greater, and more pernicious, influence on BOLD-MRI results than thought by Michaely et al. As the kidney is surrounded by adipose tissue, removal of the out-of-phase subtraction effect caused by this tissue is vital to minimizing fluctuations in T2* values. We, therefore, used in-phase echo time (TE) of the longest possible duration to limit the T2* errors and provide more accurate data, while balancing the signal-to-noise ratio. Michaely et al. used an unusually short TE of just 40 ms. We, and other similarly focused groups, understand the limitations of this novel modality and hope that the article does not discourage the application of BOLD-MRI to CKD assessment. 1. Michaely HJ, Metzger L, Haneder S et al. Renal BOLD-MRI does not reflect renal function in chronic kidney disease. Kidney Int 2012; 81: 684–689. 2. Inoue T, Kozawa E, Okada H et al. Noninvasive evaluation of kidney hypoxia and fibrosis using magnetic resonance imaging. J Am Soc Nephrol 2011; 22: 1429–1434. 3. Neugarten J. Renal BOLD-MRI and assessment for renal hypoxia. Kidney Int 2012; 81: 613–614. 4. Mimura I, Nangaku M. The suffocating kidney: tubulointerstitial hypoxia in end-stage renal disease. Nat Rev Nephrol 2010; 6: 667–678.
Tsutomu Inoue1, Eito Kozawa2, Hirokazu Okada1 and Hiromichi Suzuki1 1 Department of Nephrology, Faculty of Medicine, Saitama Medical University, Saitama, Japan and 2Department of Diagnostic Radiology, Saitama Medical University International Medical Center, Saitama, Japan Correspondence: Tsutomu Inoue, Department of Nephrology, Faculty of Medicine, Saitama Medical University, 38 Morohongo, Moroyama-machi, Iruma-gun, Saitama, Japan. E-mail: [email protected]
(CKD) with different degrees of severity of renal failure, in which blood oxygen level–dependent magnetic resonance imaging (BOLD-MRI) was used as a non-invasive technique for monitoring the degree of renal oxygenation. This study addressed the contention that intrarenal chronic hypoxia underlies the progressive nature of most forms of CKD.2–4 The authors were unable to discriminate between patients with different stages of CKD using this method. We would argue that BOLD-MRI is not the appropriate method for assessing oxygenation in CKD given the specific evolution of the pathology in this setting. BOLD-MRI measures vascular oxygen saturation with the signal being indirectly related to deoxyhemoglobin concentration.5,6 To compare different stages of disease for the degree of hypoxia, the volume of the vessels and the number of red blood cells must be the same in the voxels being compared, to make the comparison valid. The patchy involvement of the chronically diseased kidney, in which microvascular obliteration occurs in the tubule-interstitial compartment, means that the expression of a single datum point to reflect the entire cortex or medulla (as portrayed in the paper by Michaely et al.) would fail to capture this heterogeneity in the early and mid-stages of the disease. Expressing the variability of the BOLD signal intensity or R2* of pixels contained within the measured region of interest for each patient may be a more appropriate way to detect differences. Equally important, in more advanced disease, both microvasculature loss and a diminished red blood cell mass occur as CKD progresses. Hence, the source of the signal for deoxyhemoglobin concentration in a volume being measured would be diluted as the disease progresses, thereby obscuring functionally relevant local hypoxic regions and making comparisons with less advanced disease invalid. We concur with the author of the accompanying editorial that ‘limitations in the technique and execution’ lead to the conclusion that ‘the hypoxia hypothesis has not been refuted’.7 The study is important, however, for raising the need for more refined use of BOLD-MRI or for alternative methods for determining intrarenal hypoxia in CKD. 1.
2.
Kidney International (2012) 82, 934; doi:10.1038/ki.2012.282 3. 4.
Limitations of BOLD-MRI for assessment of hypoxia in chronically diseased human kidneys To the Editor: The article by Michaely et al.1 describes a sizeable population of patients with chronic kidney disease 934
5.
6.
7.
Michaely HJ, Metzger L, Haneder S et al. Renal BOLD-MRI does not reflect renal function in chronic kidney disease. Kidney Int 2012; 81: 684–689. Fine LG, Norman JT. Chronic hypoxia as a mechanism of progression of chronic kidney diseases: from hypothesis to novel therapeutics. Kidney Int 2008; 74: 867–872. Nangaku M. Chronic hypoxia and tubulointerstitial injury: a final common pathway to end-stage renal failure. J Am Soc Nephrol 2006; 17: 17–25. Long DA, Norman JT, Fine LG. Restoring the renal microvasculature to treat chronic kidney disease. Nat Rev Nephrol 2012; 8: 244–250. Mangalathu Arumana J, Li D, Dharmakumar R. Deriving blood-oxygenlevel-dependent contrast in MRI: an evaluation of T2*-weighted, T2-prepared and SSFP methods at 1.5 T and 3.0 T. Magn Reson Med 2008; 27: 1037–1045. Dharmakumar R, Qi X, Hong J et al. Detecting microcirculatory changes in blood oxygen state with steady-state free precession imaging. Magn Reson Med 2006; 55: 1372–1380. Neugarten J. Renal BOLD-MRI and assessment for renal hypoxia. Kidney Int 2012; 81: 613–614. Kidney International (2012) 82, 934–937
letter to the editor
Leon G. Fine1 and Rohan Dharmakumar1
5.
1
Department of Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California, USA Correspondence: Leon G. Fine, Department of Biomedical Sciences, Cedars-Sinai Medical Center, Los Angeles, California 90048, USA. E-mail: [email protected] Kidney International (2012) 82, 934–935; doi:10.1038/ki.2012.283
Gloviczki ML, Glockner J, Gomez SI et al. Comparison of 1.5 and 3 T BOLD MR to study oxygenation of kidney cortex and medulla in human renovascular disease. Invest Radiol 2009; 44: 566–557.
Henrik J.M. Michaely1 1
Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Mannheim, Germany Correspondence: Henrik J.M. Michaely, Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Mannheim, BA-Wu¨, Germany. E-mail: [email protected] Kidney International (2012) 82, 935; doi:10.1038/ki.2012.284
The Author Replies: In their letters to the editor, two groups, from Los Angeles1 and Saitama, Japan,2 respectively, raised some concerns regarding the methodology and the results of our previously published study on renal blood oxygen level– dependent magnetic resonance imaging (BOLD-MRI) in a non-selected patient group. The current literature on renal BOLD-MRI and its applications presents heterogeneous results regarding the correlation between R2* and the estimated glomerular filtration rate.3,4 These different results are most likely due to a lack of understanding of the complex interaction between the different types of renal diseases, their specific pathophysiology, and the complex magnetic resonance (MR) signal generation. The choice of shorter echo times in our study is surely not a relevant factor explaining the differences between our findings and the findings of other research groups. In contrast, the use of 3T MR mandates the use of shorter echo times to yield reliable image quality even with the longest echo time used.5 In addition, similar echo times as in our study have also been used by the group of Prasad et al., who pioneered the field of renal BOLD-MRI. Much more interesting than these technical details are the implications of our study that have been addressed by the authors of these letters. We concur that our study does not imply that BOLD-MRI of the kidneys should not be done. The main conclusion of our study was that ‘broad application of renal BOLD MRI seems not suitable’.3 The use of BOLD imaging in dedicated patient groups should not be discouraged. In contrast, further studies in patient groups with similar pathophysiological changes or intraindividual studies are warranted. Under these premises the influence of altered microvascular density or of the red blood cell count and altered medications could be better controlled. Novel approaches for image analysis of the parametric R2* maps, such as the suggested regionof-interest–based histogram plot, might further provide better insights into pathology. 1. Fine L, Dharmakumar R. Limitations of BOLD-MRI for assessment of hypoxia. Kidney Int 2012; 82: 934–955. 2. Inoue T, Kozawa E, Okada H et al. Is there no future for renal BOLD MRI? Kidney Int 2012; 82: 934. 3. Michaely HJ, Metzger L, Haneder S et al. Renal BOLD-MRI does not reflect renal function in chronic kidney disease. Kidney Int 2012; 81: 684–689. 4. Inoue T, Kozawa E, Okada H et al. Noninvasive evaluation of kidney hypoxia and fibrosis using magnetic resonance imaging. J Am Soc Nephrol 2011; 22: 1429–1434. Kidney International (2012) 82, 934–937
Drug use and nephrotoxicity in the intensive care unit: are we under-dosing antimicrobials in patients with acute kidney injury with the need of extended dialysis? To the Editor: An important aspect that we would like to raise, in addition to what Dr Perazella described as drug dose adjustment1 in a recent review in this journal, is the use of antibiotics in extended dialysis (ED). With ED, patients with acute kidney injury have a potentially supraphysiologic drug clearance during ED procedures, and this may be detrimental in patients for whom bactericidal concentrations of antibiotics are needed. Therefore, when the antibiotic is given in relation to ED may be as important as the dosing. This premise has been demonstrated by Lorenzen et al.,2 who reported that ampicillin/ sulbactam given too early or late in relation to ED could result in supra- or subtherapeutic concentrations. Also considered should be the pharmacodynamic properties of the antimicrobial in question. Antimicrobials require specific attention to their pharmacodynamic activity, as drugs that are time-dependent killers may require a different timing and dosing compared with those that kill in a concentration-dependent manner.3 Taking these considerations into account may be difficult enough, but complexity is added by the intensive care unit (ICU) environment, to which access is limited, and procedures may interfere with drug and dialysis scheduling. In addition, there is paucity in dosing information of several antibiotics commonly used in the ICU in patients with ED. ED pharmacokinetics has been conducted in very few drugs.4 We agree with Dr Perazella that dosing of medications in critically ill patients with acute renal replacement therapy is challenging, and that practitioners should seek best evidence and adjust their recommendation of dialysis parameters and timing to optimize the dosing of antimicrobials. 1. 2.
Perazella MA. Drug use and nephrotoxicity in the intensive care unit. Kidney Int 2012; 81: 1172–1178. Lorenzen JM, Broll M, Kaever V et al. Pharmacokinetics of ampicillin/ sulbactam in critically ill patients with acute kidney injury undergoing to extended dialysis. Clin J Am Soc Nephrol 2012; 7: 385–389.
935