June 9, 2005 - The Journal of the American Society of Nephrology (JASN) has published a JASN Express article (published on-line ahead of print copy) on the results of randomized trial that looked at the relationship between intradialytic blood volume monitoring and patient morbidity.

As both the results and control group data were unexpected and unusual, RenalWEB asked the manufacturer of the blood volume monitoring equipment, Hema Metrics, to comment on the trial results. A link to the abstract of the JASN Express article and the comments from Hema Metrics are included below:

"Intradialytic Blood Volume Monitoring in Ambulatory Hemodialysis Patients: A Randomized Trial" - abstract from JASN Express

---

Interpreting the CLIMB Study

This is an informative and unusual paper [1] indicating that prospective randomized trials are not a priori gold standard windows through which one sees immutable truths. They are simply tools. Like all tools, they can produce to flawed products. The CLIMB study shows that the validity of findings from prospective trials should be routinely subjected to formal validity evaluations and those analyses should be described to the public as part of the study report.

The analytical product of the CLIMB study indicated that both death and hospitalization risks were substantially worse if hemodialysis patients underwent routine intravascular blood volume monitoring (IBVM) than if they did not. Intermediated outcomes (like blood pressure and relative body weight) were not different between the IBVM and Control groups. IVBM is a hydration monitoring tool so one expects on clinical grounds that other indices of volume status should be worse in monitored patients if actions taken, or not taken, by clinicians in response to IBVM caused or contributed to higher mortality. They were not. How then does one evaluate this study?

A complete discussion of validity analysis exceeds the scope of these remarks. Others have developed the subject in substantial detail [2]. A complete validity analysis, including for example tests of conclusion validity, internal validity, and construct validity, was not described in the manuscript. Tests of external validity � how the findings fit into the real clinical world - were performed, however.

The mortality and hospitalization rates of both the IBVM and Control groups were compared to the parent US hemodialysis population from which the samples were drawn. The demographic attributes of both study groups were similar to those of US patients. The mortality rates in the IBVM and Control groups were 77% and 26% of the US rate respectively � both lower than in US dialysis patients but dramatically lower in the Control group. The hospitalization rate statistics were 75% and 51% of the US rate respectively � again both better than the national rate but remarkably better in the Control group (IBVM = 1.5 hospitalizations/year; Control = 1.0; US = 2.0). One could reasonably conclude, then, that IBVM improves mortality and morbidity slightly but doing nothing dramatically improves clinical outcome! That claim is said simply to make the point that the CLIMB study fails its validity test. The validity data that are presented strain the limits of logic and reason indicating that something went terribly wrong with this study. The primary findings resulting from it are therefore suspect � prospection, randomization, and control not withstanding.

One Co-principal Investigator summarized the data nicely. �It is not reasonable to conclude that an annual mortality rate of 6.4% (i.e. the Control group) can be sustained in a parent population with a rate that is over 23%. Something is wrong.� A reviewer commenting on behalf of a journal to which this manuscript was previously submitted opined that, �My biggest worry is that this is a false positive trial. �He (she) went on to enumerate the worries; 1] main outcome not anticipated, 2] no plausible link between intervention and outcome, 3] and others. The reviewer was worrying about the essential elements of a formal validity analysis. He went on to suggest an editorial if the paper were published. Another reviewer for that journal also opined that, �An editorial about the value of randomized trials and the general approach to trials may be warranted with this article.�

We agree with those sentiments that the greatest lessons taken from the CLIMB exercise concern the conduct and evaluation of clinical trials. Others have become concerned with that problem in recent years [3,4]. The CLIMB analysis is highly misleading absent its evaluation of external validity. We suggest that formal validity analyses should be a structured part of all clinical trials.


Edmund G. Lowrie, MD
Coauthor of the paper [1] and
Board member of HemaMetrics



References:

1. Reddan DN, Szczech LA, Hasselblad V, Lowrie EG, Lindsay RM, Himelfarb J, Toto RD, Stivelman J, Winchester JF, Zillman LA, Califf RM, Owen WF: A randomized trial of intradialytic blood volume monitoring in ambulatory hemodialysis patients. Jour Amer Soc Nephrology (16): 2005
   
2. Trochim WMK: The research methods knowledge base (Second Edition). Atomic Dog Publishing, Cincinnati, OH, 2001. ISBN 0-9701385-9-8.
   
3. Kapatchuck TJ: The double-blind, randomized, placebo-controlled trial: Gold standard or golden calf. Jour Clin Epidemiol 54: 541-549; 2001
   
4. Miettinin OS, Yankelevitz DF, Henschke CI: Evaluation of screening for cancer: annotated catechism of the Gold Standard creed. Jour Eval Clin Practice 9: 145-150; 2003
   

---