Clinical judgment is the most important element in overhydration assessment of chronic hemodialysis patients

While the assessment of solute clearance has moved forward substantially in recent years, the estimation of adequate fluid removal remains a challenging problem in the management of hemodialysis (HD) patients. Dialysis-associated overhydration (OH) and dehydration have been linked to adverse events. Chronic OH is a major factor in the development of arterial hypertension, although the causal relationship between OH, hypertension and mortality is intricate due to the higher prevalence of comorbid conditions in HD patients. Hypotension resulting from excessive ultrafiltration can provoke acute ischemic events with recurrent episodes, potentially causing functional impairment and organ damage [1, 2].

Dry weight (DW) has been conventionally defined as the lowest weight that can be tolerated without developing symptoms of hypovolemia. Although based on trial and error, probing for DW has been a common practice. Today, it is not simply a symptom-guided probing anymore, but rather a complex systematic clinical approach, including laboratory data and imaging techniques. Patient-reported symptoms can be misleading without knowing the medical history and usually become more specific as the OH increases [3]. Patients differ in autonomic system responsiveness, vascular refilling capacity, comorbidities and their therapy. Advanced kidney disease is accompanied by metabolic alterations, often resulting in decrease in body cell mass, increase in extracellular volume and consequently OH. Body composition undergoes changes yet again after a patient starts HD treatment and the uremic environment improves. All this together makes an accurate assessment of hydration in HD patients very challenging. The role of many laboratory and imaging parameters as potential indicators of OH has been evaluated, but with inconsistent results. Bioelectrical impedance (bioimpedance, BIA) offers the possibility of direct measurement of extracellular and intracellular fluid compartments [4]. It gained more attention in recent years when several studies reported superiority of BIA in the assessment of dialysis OH [5]. Unfortunately, with the introduction of new technologies, there has been an indisputable tendency to undervalue the significance of clinical judgment in hydration status estimation.

The objective of the present study was to evaluate the relevance of clinical judgment in the assessment of pre-HD OH. To accomplish this, we compared the performance of three different methods of OH estimation: (1) clinical judgment guided by a single clinical examination with (2) multifrequency bioimpedance analysis and (3) complex systematic clinical approach. We additionally examined the associations of these methods with selected laboratory and imaging parameters.

An optimal method should have high sensitivity and specificity, while still being generally applicable and cost-effective. The systematic clinical approach is a system combining physician and patient inputs, laboratory data and imaging. Clinical judgment guided by clinical examination is a crucial component of the systematic clinical approach. Our models have identified clinical judgment as the single most important factor in OH assessment. BIA reliably measures ECW and calculates OH_BIA using a body composition model, based on reference data obtained from the normal population. Dry weight determined from the computerized OH_BIA cannot be always applied and achieved without the risk of dehydration and, therefore, does not represent the optimal DW in every patient. The discrepancy between measured ECW and calculated OH_BIA can be at least partly explained by not implementing the option for patient-individualized and/or disease-specific correction. Hence, the general applicability of the computed OH_BIA remains uncertain. We see a particular strength of BIA in the direct estimation of ICW that defines body cell mass and cannot be reliably assessed by other routine techniques. Malnutrition, a commonly undiagnosed condition in dialysis patients, leads to loss of lean body substance [7]. Implementation of serial ICW measurements in individual patients would be able to unmask a clinically inapparent decline in body mass, prevent an increase of OH, and uncover an underlying process possibly requiring further medical intervention. This interpretation is supported by our models, which selected ICW as the most significant BIA parameter in OH assessment. Our analyses make evident that only combinations of several methods and parameters provide an acceptable prediction precision. The integrative function of clinical judgment is reflected by the better accuracy of models with implementation of OH_CLI and also by the highest predictive importance of OH_CLI.

Despite similar hydration characteristics, our patients had lower BP than the study subjects reported by Chazot [8]. However, many studies do not report antihypertensive drugs prescribed only for cardioprotection, which creates inconsistency. We think that this different indication does not eliminate the antihypertensive effect, and included them in our analysis. Investigators from Tassin in France described patients who remain normotensive despite being above calculated DW, and explain this by better clearance of vasoactive substances during the long HD practiced in the Tassin dialysis center [9]. Our patients presented a normal average BP that correlated with OH. This emphasizes that BP changes rather than absolute values in individual patients, even within normal limits, may be indicative of OH. Undetected overhydration, silent hypervolemia, may result in hypertension as late as 12 h after leaving HD [10]. For this reason, we believe that regularly performed 24-h BP monitoring should be a standard component of hydration evaluation in HD patients.

The calf has a relatively uniform structure with better hydration, and recent evidence has suggested that calf BIA may be more sensitive than the whole-body method [11]. We could prove a strong link between calf circumference and OH parameters, and provide further support for this emerging technique. The conventional indicators of volume overload in the non-HD population, chest X-ray or echocardiography, might not be that reliable in HD patients. Fluid oscillations associated with HD can induce organ remodeling (atrial dilatation, ventricular hypertrophy, increased pulmonary vascular resistance), and decrease the specificity and sensitivity of these techniques for fluid overload. Nevertheless, compared to BIA, conventional imaging still offers additional information that can be valuable for treatment decisions. VCD and collapsibility variations have been reported as sensitive indicators of OH, but the recommended interval of at least 1 h after HD limits the use of VC sonography in ambulatory patients [12]. Our models showed a high predictive role for VCCI in OH estimation (second best after OH_CLI), also before HD. There are only a few studies examining the effects of HD on pulmonary functional parameters. The importance of spirometry in OH assessment has not been studied so far, and our data indicate rather an inferior role in HD.

It is evident that any of the single parameters is accurate enough to predict the extent of OH by itself. Clinical judgment of an experienced physician was the single most significant element in OH assessment, and showed the highest predictive value in combination with other variables as well. Admittedly, clinical judgment is observer-dependent and difficult to standardize. Nevertheless, the non-standardized decision choice is precisely the unique feature of clinical judgment. Studies examining different approaches to OH assessment in large patient populations typically report only the average values of the accuracy, without correlations to their standard method, which obscures the performance in individual patients. We need a method that can be applied and remains precise and reliable also in smaller groups of patients, typically encountered by dialysis physicians in routine clinical practice. Our study demonstrated that a combination of integrative clinical judgment with routine techniques is a precise and valuable tool in hydration status assessment in HD patients. BIA, a quick, reproducible and non-invasive bedside measurement, may help to identify changes in body compartments not fully appreciated by clinical or biochemical assessment. However, the most important question, whether the improved accuracy of OH assessment resulting from implementation of technological advances will also be reflected in improved patient outcomes, requires further investigation.