Response to “Influence of Industry on Renal Guideline Development”

Dr. Coyne argues (1) that our statement 2.1.2, Upper Limit of Hemoglobin (1), will promote greater erythropoiesis-stimulating agent (ESA) use, enrich pharmaceutical companies and dialysis providers, and impoverish the health care system; (2) that the upper limit should have been recommended on the basis of the Food and Drug Administration (FDA) package insert and unpublished data; (3) that pharmaceutical industry support for the National Kidney Foundation (NKF) and industry ties among NKF Kidney Disease Outcomes Quality Initiative (KDOQI) membership foretold the upper limit statement; (4) that shame and discredit will befall the renal community as a result; and (5) that the only solution is to preclude pharmaceutical support and experts with conflicts of interest (COI) from guideline development and shift guideline development responsibility to the National Institutes of Health. Although some of Dr. Coyne’s suggestions deserve support, others warrant scrutiny:

• Dr. Coyne objects to the specific value of 13.0 g/dl as the upper hemoglobin “target” and offers his own alternative. However, the actual statement 2.1.2 reads, “There is insufficient evidence to support routinely maintaining hemoglobin levels at 13.0 g/dl or greater in ESA-treated patients.” That statement defines strength of evidence, not target hemoglobin. The carefully selected words are supported in exhaustive detail by 21 pages of tables, figures, and rationale. Dr. Coyne’s contention that KDOQI raised the target hemoglobin represents a fundamental misinterpretation of the guideline.

• Dr. Coyne misrepresents both statements 2.1.2 (Upper Limit of Hemoglobin) and 3.2.4 (Upper Level of Ferritin). He dismisses the two as “opinion-based” and therefore susceptible to industry influence. Each statement, however, is supported extensively in the rationale with an analysis of the available evidence. The terms clinical practice guidelines (CPG) and clinical practice recommendations (CPR) distinguish not between the presence and the absence of evidence but between levels of the strength and quality of evidence.

• Dr. Coyne assumes that ESA use is tied to the upper limit of hemoglobin. In fact, ESA use is determined largely by treatment protocols, population-based benchmarks, and reimbursement policy, not by what is targeted as the upper limit of hemoglobin in an individual patient or, for that matter, mean national hemoglobin. ESA use in the United States far exceeds that of any other country, independent of mean national hemoglobin (Figure 1) (2). This is an issue of public policy and implementation, not guideline development.

• The statement that “DaVita has increased its hemoglobin target to 12.5 to 13.0 g/dl” is false. DaVita policy calls for maintaining hemoglobin ≥11.0 g/dl and decreasing epoetin dosage if hemoglobin exceeds 12.0 g/dl.

• The statement, “The chair and co-chair define the scope of work and the specific questions to be addressed,” although factually correct, is taken out of context. Co-chairs prepare the first draft of the scope-of-work document as a series of open-ended questions to be considered by work group members. At their first 2-d meeting, members add questions until the initial working document is considered complete. The inclusive, combined set of questions (still without answers) forms the basis for the deliberation and discussion that follows and evolves as the group works, serving to ensure that all topics that are deemed clinically relevant and worthy of review are identified and addressed. Absence of a topic from the final guideline should not be regarded as evidence that the pertinent question was not identified or addressed by the work group.

• The FDA product prescribing information (FDA labeling) is not a guideline. As a policy or implementation tool, it potentially is confusing. Current labeling for ESA products that are available in the United States includes reference to a “suggested target” hemoglobin range of 10 to 12 g/dl for epoetin alfa (3,4) but specifies no range for darbepoetin (5), just a warning that the hemoglobin should not exceed 12 g/dl (3–5). Finally, the labeling for two of the three agents includes the statement, “At the physician’s discretion, the suggested target hemoglobin range may be expanded to achieve maximal patient benefit” (3,4). The labeling offers no evidence to support the specific upper hemoglobin target, affords no explanation for the differences in hemoglobin target definition among the three ESA products, and cites no CKD literature that was published later than 1998 (3–5).

• The Anemia Work Group, like all KDOQI work groups and most other guideline development efforts, required a priori that information that is used as evidence for the guidelines had been peer reviewed and either published or accepted for publication. This requirement served as a safeguard against bias from inclusion of unpublished, unreviewed results from trials that were not fully adjudicated (Correction of Hemoglobin and Outcomes in Renal Insufficiency [CHOIR]) or not actually completed (Dialysis Patients Response to IV Iron with Elevated Ferritin [DRIVE]).

• Dr. Coyne’s request to have transparency in disclosure and a process to resolve issues around real or perceived COI is well taken. However, the proposal that guideline development be carried out only by “intelligent physician-scientists with no COI,” excluding both pharmaceutical support and experts with potential COI, although attractive in theory, is unrealistic. As we have pointed out, the KDOQI guideline development process includes an Evidence Review Team without COI, designed to provide methodologic expertise to the work group and to ensure that problems are defined clearly, data are interpreted appropriately, and rigor and transparency are maintained. If experts with COI were to be excluded from guideline development as unfit to appraise evidence, then the same standard should apply to all those who are in a position to appraise evidence, including each author who submits a trial report to a journal, each member of each journal’s editorial board, and each person to whom a manuscript is submitted for peer review. Absolute proscription of COI, if achievable, would discourage investigators from participating in the design, conduct, safety monitoring, and analysis of industry-sponsored research. The likely result, less clinical research that yields evidence of lower quality, would be undesirable. Guideline development aims not to eliminate but to control all potential or perceived bias, including that arising from COI. The goal is to control potential bias by adhering to scientific and methodologic rigor; fostering an interdisciplinary and independent work group; and using an open, structured, and systematic development process.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Relationship between mean hemoglobin and mean weekly epoetin dosage for hemodialysis patients in 12 countries. Data from 2002 to 2003. Adapted with permission from reference (2).

In conclusion, we disagree with Dr. Coyne’s representation of key guidelines, some of the facts that he has used to support his arguments, and his conclusion. We entirely agree that guideline development should follow and document a transparent process for managing potential or perceived COI in work groups, and we urge the NKF to remain at the forefront in developing new methods to control work group bias in all forms. We share Dr. Coyne’s conviction that “KDOQI guidelines have been a tremendous asset to the renal community and patients with CKD alike. The improvement and standardization in care as a result of KDOQI cannot be overstated.” Finally, we are justly proud of the work of all members of the Anemia Work Group and of their achievement, the KDOQI 2006 Anemia Guidelines (1). For the countless hours that the work group served on behalf of the nephrology community, we are deeply grateful.

• Copyright © 2007 by the American Society of Nephrology