Hemoglobin Variability in Dialysis Patients

Hemoglobin variability is the extent to which multiple measured hemoglobin values differ from each other. Variability may be assessed within the same patient or between patients in a group; in the context of clinical practice, it is generally the variability within a patient that is important, whereas for quality assurance purposes, both variability within patients (an index of individual stability) and between patients (an index of the extent to which values differ between patients) may be relevant. As West et al. observe in the current issue of JASN,1 the adjustment of epoetins in the management of anemia in renal disease, whether done by clinical judgment, the use of simple clinical decision rules, or a more complex computer program essentially follows the principles of a negative-feedback loop; that is, a derangement leads to a dose adjustment in the direction predicted to bring the patient's hemoglobin back toward the desired value or into the desired range. This mechanism means that instability and, in some cases, a degree of periodicity is an inherent and inevitable feature of the system. Different methods have been used to quantify the degree of variability. West et al. use the absolute value of the rate of hemoglobin change (calculated from curve-fitting computer algorithms), which they call the trajectory, measured in g/dl per mo.1 Based on individual curve fitting, it is applicable only to the assessment of within-patient variability, although, as they have done, these values can then be aggregated and compared between groups using standard statistical techniques. Other measures of variability that can be assessed within a patient or across a group of patients are the SD or the coefficient of variation (the ratio of the SD to the mean). Finally, the proportion of time outside certain thresholds can be assessed on the basis of either actual hemoglobin measurement or rolling averages of hemoglobin measurements. Targets may be defined for a number of purposes. First, a target might be defined from basic and clinical science data to encompass the values thought to be associated with the optimal combination of quality and length of life. Second, clinical decision rules or algorithms often set target ranges pragmatically as a range of values within which no dose adjustment is necessary. And third, target ranges may be used by individuals, groups,2 or by society3 to assess the efficacy of treatment in meeting specified goals. We suggest that these three purposes are quite distinct, and the optimal target and range for each may differ. Reasons for variability include abrupt changes caused by distinct comorbid events such as bleeding or transfusion. In addition, chronic comorbidity (particularly inflammation), iron stores, dialysis adequacy, water quality, residual renal function, hyper- or hypoparathyroidism, B12 or folate deficiencies, seasonal effects, the use of angiotensin-converting enzyme inhibitors and possibly angiotensin receptor–blocking drugs, and inherent, currently unmeasurable patient-specific factors all may lead to variability between patients. Changes in these factors would lead to increased variability within an individual patient over time. The current work by West et al. suggests a new metric: The sum of these factors may reflect the sensitivity of the patient. Changes in volume status and unavoidable sampling and laboratory measurement errors lead to further variability. Finally, the frequency of measurement, frequency of dose adjustment, frequency of dosing, and pharmacokinetics of the epoetin used are important further factors that, even in a perfectly stable situation, affect the amplitude and periodicity of the hemoglobin trajectory. West et al. have used a novel methodology to assess within-patient variability. Individual patients’ hemoglobin values are plotted and curves fitted that pass through or near data points. This permits the calculation of the slope, or rate of hemoglobin change, a value that changes instantaneously. The average of this value assesses an individual patient's variability. Plotting rate of hemoglobin change against the absolute value for an individual patient allows graphical interpretation in that tighter ellipses are indicative of better control. This offers a new methodology for assessment of variability in future studies. Variability has previously been shown to be increased in patients who are younger, have lower albumin and higher serum ferritin (likely because these last are inflammatory markers), and have higher mean corpuscular hemoglobin.4 Important unanswered questions in this area relate to modifiable variables: The optimal frequency of measurement, frequency of dose adjustment and magnitude of dose increments in unselected patients receiving specific epoetins, and optimal iron protocols. Iron-loading strategies may cause more abrupt increases in hemoglobin than iron-maintenance protocols.5 It is probable that longer-acting agents lead to greater stability at a given dose frequency—under the experimental conditions used in the current paper, stability was greater with a longer-acting epoietin compared with a shorter-acting agent.1,6 The width of the target range may also affect variability, but empiric data here are confusing. Two previous randomized controlled trials conducted by Will's group in the same population of patients compared a target range of 10.5 to 14 g/dl with a range of 11.5 to 14 g/dl, and a target range of 11 to 12 g/dl with a range of 11 to 13 g/dl.7,8 In the first study, no statistically significant reduction in group SD resulted from the narrower target range; however, in the second study a statistically significant reduction occurred in the group managed with the narrower target.7,8 This second study was also of interest as an example of a difference between thresholds for intervention and the thresholds used as a measure of success. The authors argued that to maintain hemoglobin above 10 g/dl in a large proportion of patients, the dose must be changed proactively as the hemoglobin crosses a threshold that is higher than this.7 Why is it important to maximize hemoglobin stability? In the management of anemia with epoetins, physicians steer individual patients between the Scylla of increased mortality caused by higher hemoglobin targets9,10 and the Charybdis of symptoms and lower quality of life from severe anemia11—increased variability reflects an increased probability of patients veering toward one of these hazards. Increased variability will also increase the proportion of patients outside given targets at a particular time or over a period of time, leading to poor performance in meeting audit targets; in some countries, exceeding hemoglobin ceilings is undesirable per se because of funding implications.3 Studies of the relationship between variability and clinical outcomes are however, to our knowledge, lacking. Research in this area is of more than technical interest. Given the high costs of epoetins, more information on cost-effective methods to maximize hemoglobin stability and clinical benefits is needed. We further suggest that future trials on any issue in anemia management consistently report the between- and within-patient SD and the statistical significance of any differences, especially in studies evaluating extended erythropoietin dosing strategies. Algorithms that take into account the current hemoglobin trajectory or trend,6 as well as the most recent value or rolling average of values, as used by West et al. in this issue, appear particularly worthy of further investigation. Neural networks also have shown some promise in this area and require further testing in clinical practice.12 DISCLOSURES C.M. Clase served on the advisory board for Hoffman-La Roche (1998). K.S. Brimble received study funding from Janssen-Ortho (2001) and served on the advisory board for Janssen-Ortho (2002).

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.