Adverse Effects Reporting Standards NEEDED


The Need for Adverse Effects Reporting Standards in Oncology Clinical Trials

Andy Trotti

H. Lee Moffitt Cancer Center, University of South Florida, Tampa, FL

Søren M. Bentzen Gray Cancer Institute, Northwood, Middlesex, UK

In this issue of the Journal of Clinical Oncology, the Groupe d'Oncologie Radiothérapie Tête et Cou (GORTEC) report long-term survival and late effects from a pivotal randomized trial of chemoradiotherapy in advanced head and neck cancer [1].

Although overall survival remains poor in this population, the therapeutic advantage of combined modality is significant and is sustained beyond 5 years.

This article adds to other phase III trials supporting the superiority of concurrent chemoradiotherapy over radiotherapy alone in head and neck cancer [2-9] and represents a real success story in clinical research.

Combined modality approaches are associated with higher acute toxicity, requiring increased levels of supportive care.

However, there are also growing concerns about the high rate of late effects, as well as acute toxicity, which may result in long-term consequential effects such as severe swallowing dysfunction and mucosal injuries.

These higher rates of adverse effects have been generally perceived as "worth it" by the oncology community. But how do we really know that the benefits outweigh the harms of these aggressive programs?

The truth is that we do not have sufficient information on neither the acute nor, in particular, the late adverse outcomes to make this judgment, making it difficult to properly compare treatments and describe risk in the informed consent process.

There are now at least six positive concurrent chemoradiotherapy trials in head and neck cancer, each employing different drug and fractionation combinations. Which is the "best" one for my patient?

If I want to consider relative rates of adverse effects among these regimens, are there accepted methods for comparing toxicity profiles?

The GORTEC trial reports a 30% risk of high-grade [3,4] late effects at 5 years in the radiation-alone arm compared with 56% in the chemoradiotherapy arm, utilizing the Radiation Therapy Oncology Group/European Organization for Research and Treatment of Cancer (RTOG/EORTC) [10] grading system for late effects, and the National Cancer Institute Common Toxicity Criteria (NCI/CTC) [11] for selected late effects not covered by the RTOG/EORTC system.

The GORTEC group has also reported a more in-depth evaluation of late toxicity in the same trial [12]. In an effort to compare the utility of different late effects grading systems, forty-four 5-year survivors were systematically screened using the individual criteria scales of three separate grading instruments: the RTOG/EORTC system, the NCI/CTC, and the Subjective, Objective, Management, Analytic (SOMA) System [13].

They found that each tool covered some nonoverlapping toxicity items. The SOMA system seemed to be the most comprehensive and perhaps the best for grading fibrosis.

The rate of late effects, when combining the data from the three tools, was 47% in the control arm and 82% in the experimental arm, which was statistically significant (P = .02), in contrast to the difference observed using the RTOG scale alone.

The original (1999) paper from the GORTEC study reported rates of 9% and 14% for grade 3 to 4 late effects in the radiotherapy (RT) and the chemoradiotherapy (CRT) arms, respectively [14].

These differences were not statistically significant. The median observation time was 35 months, but these incidence estimates were crude ratios between the number of patients with late effects and the total number of patients treated.

Actuarial estimates, typically employing the Kaplan-Meier method, adjust for the number of cases under observation at the time when late effects become clinically manifest, providing a more accurate determination of the prevalence of late effects in long-term survivors.

The importance of establishing reporting standards is illustrated by the large range of reported values (even considering some increase in the crude estimate of late effects between years 3 and 5), from 9% to 47% in the RT arm and from 14% to 82% grade 3 to 4 late toxicity in the CRT arm in the three reports from the GORTEC trial.

The GORTEC group deserves credit for their systematic attempt to improve the assessment and reporting of late effects. At the same time, their study shows that the mere choice of a grading instrument and reporting methodology can dramatically alter the observed rates of adverse effects.

This raises tough questions regarding the validity of the RTOG/EORTC scale and other instruments, as well our ability to compare toxicity outcomes among trials and institutions.

Moreover, if we believe that 50% to 80% of survivors suffer severe to life-threatening long-term toxicity, should we consider this an acceptable regimen?

While there have been efforts to report adverse effects in oncology clinical trials for at least six decades, there have been extreme variations in the methods of reporting, and little attention to the development of reporting standards or new methods of reporting [15-20].

Current methods of summarizing and communicating adverse effects are largely anchored in safety reporting methods for single-agent nononcology products. The use of simple adverse effects incidence tables is much too crude to characterize the severity, time-related multimodality, and cumulative aspects of cancer treatment toxicity.

There has been little organized effort in oncology to develop more advanced reporting methods to capture the full scope of these clinically observed adverse effects.

Quality-of-life (QOL) investigators have tried to reflect the impact of adverse effects utilizing patient-rated instruments. QOL is fundamentally a subjective measure, significantly modulated by one's ability to adapt to adversity. Q

OL and toxicity are strongly correlated in some studies, and not in others where one would expect it to be. QOL includes the diverse domains of social support and spirituality. The clinical utility of QOL instruments remains elusive.

QOL (assessed by patients) may be complementary to adverse effects reporting (assessed by clinicians), but it is not a substitute [19,20].

For cooperative group trials in oncology, toxicity grading systems are used to facilitate the recognition and severity ranking of adverse effects.

For the last 20 years, the dominant grading systems have been the WHO and the NCI/CTC systems for reporting acute toxicity, although several group-specific and ad hoc systems have also been used [21,22].

Late effects systems include the RTOG/EORTC, SOMA, and Dische systems [23]. By nature, late effects are more difficult to recognize and grade. Grading systems vary in modality focus and severity scaling, such that a low-grade reaction in one system may be considered high-grade in another.

For acute toxicity, reporting practices usually include a table of the incidence of hematologic and nonhematologic events containing a variable number of toxicity items.

There are variations in methods of display, including different combinations of grades and ad hoc summary methods to add up all toxicity items, including the "worst grade method" [24].

Publications rarely describe these methods in any detail. The reporting of late effects is much less common than acute effects, with even wider variations in grading systems, frequency of evaluation, intensity of screening, and methods of analysis.

The reporting of late effects in single modality radiotherapy trials is more common than in combined modality trials, reflecting a more developed awareness of late effects issues among radiation oncologists [15].

Table 1 provides a snapshot of toxicity reporting among nine frequently cited combined modality head and neck cancer trials published within the last 10 years, including the GORTEC report.

While all contain some kind of acute toxicity reporting, the methods vary considerably, including four different recognized grading systems and two descriptive efforts. There was a range in the number (3-10) and types of items.

Only two items are common to all nine reports: neutropenia and mucositis/stomatitis. All of these trials have a median follow-up of at least 24 months, which is ample time to express a significant proportion of late effects.

Only four of nine used a recognized grading system, with two providing descriptive information, and three with no report of late effects. Summary rates of high-grade late effects reported in four trials in the control and experimental arms were, respectively, 15% and 20% (Brizel et al–calculated [5]), 6.4% and 10% (Wendt et al [7]), 34% and 58% (Jeremic et al–calculated [9]), and 30% and 56% using two instruments, and 47% and 82% using three instruments (GORTEC [1]).

Only one trial reported on long-term swallowing function (Starr et al [8]). The risk of feeding tube–dependence beyond 2 years was disturbingly high: 25% in the control arm and 50% in the experimental arm (P = .02). No trials reported on laryngeal function or tracheostomy tube dependence.

In terms of the quality of toxicity reporting, the GORTEC trial would rank high. But even this effort falls short of providing an adequate picture of late effects, in that only 44 of 326 patients were evaluated.

Crude prevalence rates were reported, not actuarial methods, which are designed to adjust for patient loss to competing risks. Despite the use of three separate grading instruments, no data were collected on the critical items of swallowing or airway function.

These gaps, inconsistencies and variations in reporting practices indicate that published toxicity reports are not only frequently lacking key information, but also suggest that they likely contain significant underreporting, bias, and errors.

There are no guidelines regarding the clinical application of a given toxicity grading system, such as methods of patient screening or data collection, and no uniform conventions for data display, analysis, or publication.

This is in sharp contract to our ability to describe prognosis, which includes multiple well-defined survival and tumor control end points quantified using standardized analytic methods.

The US Food and Drug Administration is in the process of developing written guidance for safety reporting in new drug applications [26]. It tries to address some of the complexities of adverse effects in oncology, but it does not provide oncology-specific recommendations.

It includes a general discussion of various approaches to eliciting data from patients and selection of adverse events for a toxicity profile table. Regarding severity and recurring toxicities, the draft document states, "It may be useful to consider displays that distinguish between events on the basis of severity" and that "no distinction is made between a patient who has one...or multiple (adverse) occurrences."

We currently have limited guidance from the NCI. The CTC Manual describes the rules for grading individual criteria and the regulations for expedited adverse event reporting, but does not address summary methods of data display [25].

In June 2003, the NCI announced the third revision of the CTC, relabeled CTCAE v3.0 (Common Terminology Criteria for Adverse Events Version 3.0) [25]. The principal changes, compared with version 2.0, are the inclusion of a full set of late effects criteria, expansion of criteria for surgical effects and improved coverage of pediatric issues. CTCAE v3.0 represents the first comprehensive multimodality grading system for reporting both acute and late effects in oncology [28].

We now have a single "grading dictionary," and a common "yardstick" for measuring severity. New terms and language will be added as new agents and adverse effects are recognized. However, even widespread adoption of CTCAE v3.0 will not assure complete and meaningful toxicity reporting.

The CTC does not address patient screening, data collection, data display, methods of analysis, or reporting guidelines.

The Consolidated Standards of Reporting Trials agreement for improving the quality of reporting in phase III trials provides some general guidance on presenting "estimates of the frequency of the main adverse events" in each group, and "the number of adverse events" when participants experience an adverse event "more than once," but specifics are lacking [29].

An analysis of the quality of adverse event reporting in AIDS and six other serious illnesses showed less than one-third of the reports to be adequate [32].

The authors called for similar investigations in other medical specialties with high toxicity profiles and the development of vigorous toxicity reporting standards [33].

We clearly lack sufficient methods to effectively communicate the full extent of the potential and expected adverse effects of cancer treatment.

The high toxicity burdens generated by oncology treatments suggest we need more advanced methods and standards than in nononcology disciplines to describe the scope and severity of adverse effects.

The need for new toxicity reporting standards in oncology is of increasing importance as decades of cancer research is resulting in an increasing number of treatment options. A host of new molecules are currently in development, many of which seem to have low toxicity profiles.

However, in many cases, these new agents are being combined with traditional cytotoxic agents and radiation.

How do we monitor for unexpected interactions among agents or compare the risks and benefits of new combinations versus traditional combinations?

Relative therapeutic benefit is easier to judge when you only have a few choices of treatment, and when those therapies and outcomes are clearly different.

Making choices among a larger number of more similar treatment options requires more definitive and interpretable information regarding adverse effects.

We need to develop and implement more meaningful and reliable toxicity reporting methods in oncology. This will require a close examination of the entire toxicity reporting process, including trial design, assessment of toxicity, data display, analysis, and reporting methods.

This will also require the coordinated planning and collaboration of multiple organizations, industries, regulatory agencies, and cooperative groups.

Will such an effort be worth it? The benefits to our patients and the oncology community will be the ability to better understand, communicate, and prevent adverse effects and will also greatly facilitate decision making in the selection of cancer therapies. The only harm is in not trying.

Authors' Disclosures of Potential Conflicts of Interest:

The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation.

Acted as a consultant within the last 2 years: Andy Trotti, Sanofi, Rx Kinetics, Amgen. Received more than $2,000 from a company for either of the last 2 years: Andy Trotti, Sanofi, Rx Kinetics, Amgen.

Journal of Clinical Oncology, Vol 22, No 1 (January 1), 2004: pp. 19-22

References-Reporting Standards

Editorial, JCO, 1/04

Quality of Published Adveres Drug Event Reports

Ann Pharmacother, 12/03

Remember we are NOT Doctors and have NO medical training.

This site is like an Encyclopedia - there are many pages, many links on many topics.

Support our work with any size DONATION - see left side of any page - for how to donate. You can help raise awareness of CAM.