Consort-uat.koneka.com

Better Reporting of Harms in Randomized Trials: An Extension of the
CONSORT Statement
John P.A. Ioannidis, MD; Stephen J.W. Evans, MSc; Peter C. Gøtzsche, MD, DrMedSci; Robert T. O’Neill, PhD; Douglas G. Altman, DSc;
Kenneth Schulz, PhD; and David Moher, PhD, for the CONSORT Group*
In response to overwhelming evidence and the consequences of
examples to highlight specific aspects of proper reporting.
poor-quality reporting of randomized, controlled trials (RCTs),
We hope that this document, in conjunction with other
many medical journals and editorial groups have now endorsed
CONSORT-related materials (www.consort-statement.org), will
the CONSORT (Consolidated Standards of Reporting Trials) state-
help authors improve their reporting of harms-related data from
ment, a 22-item checklist and flow diagram. Because CONSORT
RCTs. Better reporting will help readers critically appraise and
primarily aimed at improving the quality of reporting of efficacy,
interpret trial results. Journals can support this goal by revising
only 1 checklist item specifically addressed the reporting of safety.
Instructions to Authors so that they refer authors to this doc-
Considerable evidence suggests that reporting of harms-
related data from RCTs also needs improvement. Members of
the CONSORT Group, including journal editors and scientists,
Ann Intern Med. 2004;141:781-788.
www.annals.org
met in Montebello, Quebec, Canada, in May 2003 to address
For author affiliations, see end of text.
this problem. The result is the following document: the stan-
For definitions of terms, see Glossary.
dard CONSORT checklist with 10 new recommendations about
*For a list of members of the CONSORT Group, see Appendix 1, available at reporting harms-related issues, accompanying explanation, and
Reporting harms may cause more trouble and discredit than the interventions may cause. We encourage authors to use the fame and glory associated with successful reporting of benefits term “harms” instead of “safety.” In addition to misused terminology, reporting of harms in RCTs has received less The CONSORT (Consolidated Standards of Reporting attention than reporting of efficacy and effectiveness and is
Trials) statement, a checklist (Table 1) flow diagram
often inadequate (6 –14). In short, both scientific evidence first published in 1996 and revised 5 years later (2, 3), is an and ethical necessity call for action to improve the quality effort to standardize, and thereby improve, published re- of reporting of harms in RCTs (15, 16). Here, we present ports of randomized, controlled trials (RCTs). One of the a set of recommendations and accompanying explanations additions to the 2001 revision was an item about reporting for the proper reporting of harms in RCTs. These recom- adverse events. This single item did not do full justice to mendations should complement the existing CONSORT the importance of harms-related issues. The CONSORT statement (Table 2). Examples are presented on the Annals
Group met in September 2001 to discuss how to correct (www.annals.org) and CONSORT (www.consort-statement this deficiency. We aimed to provide evidence-based guid- ance on the reporting of harms in RCTs. First, we searchedMEDLINE, EMBASE, Web of Science, and the CochraneLibrary using a wide array of terms related to harms and RECOMMENDATIONS
identified pertinent evidence. We also communicated with Title and Abstract
experts and reviewed bibliographies of identified articles to Recommendation 1. If the study collected data on harms find additional studies. At a meeting in Montebello, Que- and benefits, the title or abstract should so state. bec, Canada, in May 2003, CONSORT Group members, The title should mention harms if the study of harms including several journal editors and additional experts in was a key trial objective. Many phase I and phase II trials, related fields, held a structured discussion of recommenda- some phase II/III trials, and most phase IV trials (17, 18) tions about reporting of harms-related issues in RCTs. The target harms as primary outcomes. Yet, the title and ab- discussions led to a written document that we circulated stract seldom contain the word “harm.” Among 375 143 among the team members for comment. The present entries in the Cochrane Central Register of Controlled Tri- manuscript describes our recommendations on the appro- als (Cochrane Library, issue 3, 2003), searching titles with the search terms harm or harms yielded 337 references The terminology of harms-related issues in RCTs is (compared with 55 374 for efficacy and 23 415 for safety).
confusing and often misleading or misused (see Glossary) Of the 337, excluding several irrelevant articles on self- (4, 5). “Safety” is a reassuring term that may obscure the harm or harm reduction, only 3 trial reports and 2 ab- real and potentially major “harms” that drugs and other stracts contained the word “harm” in their titles.
Improving Patient Care is a special section within Annals supported in part by the U.S. Department of Health and Human Services (HHS) Agency for Healthcare Research and Quality(AHRQ). The opinions expressed in this article are those of the authors and do not represent the position or endorsement of AHRQ or HHS.
2004 American College of Physicians 781
Improving Patient Care Improving the Reporting of Harms Glossary
Introduction
Background
Adverse events: Side effects that are harmful. However, side effects suggest Recommendation 2. If the trial addresses both harms and causality (effects caused by the tested intervention). Some authors usethe term “adverse effects” synonymously with “side effects”. In the benefits, the introduction should so state. typical randomized trial, it is difficult to know whether an observed event The introduction states the scientific background and is partially or entirely due to the intervention or whether it is totallyunrelated to the intervention (for example, a consequence of the rationale of an RCT. This requires a balanced presentation underlying disease process). The purpose of a trial is to collect and whereby the possible benefits of the intervention under appropriately report good and bad events and outcomes so that they may investigation are outlined along with the possible harms be compared across treatment groups. In this regard, the term “adverseevents” is probably better to describe harmful events that occur during associated with the treatment. Randomized, controlled tri- als that focus primarily on harms should clearly state thisinterest when describing the study objectives in the Intro- Adverse reaction and adverse drug reaction (ADR): Events for which a causality link to the tested intervention is well established and strong duction and in defining these objectives in the Methods.
enough (sensitive and specific) to warrant attribution of the event to theintervention (for details, see definitions proposed in references 4 and 5).
Attribution of causality in the setting of clinical trials may be misleading.
Outcomes
Harms: The totality of possible adverse consequences of an intervention or Recommendation 3. List addressed adverse events with therapy; they are the direct opposite of benefits, against which they must definitions for each (with attention, when relevant, to grading, expected vs. unexpected events, reference to standardized and Passive surveillance of harms: The recorded adverse events are those that validated definitions, and description of new definitions). the study participants spontaneously report on their own initiative. In The Methods section should succinctly define the re- active surveillance of harms, participants are asked about the occurrenceof specific adverse events in structured questionnaires or interviews or corded adverse events (clinical and laboratory). Authors predefined laboratory or other diagnostic tests are performed at should clarify whether the reported adverse events encom- pass all the recorded adverse events or a selected sample.
Risk–benefit ratio: The most common expression for the comparison of They should explain how, why, and who selected adverse harms and benefits. It is a technical term that assumes that a ratio can events for reporting. In trials that do not mention harms- indeed be calculated. Because the benefits and harms of an interventionare often so different in character or are measured on different scales, the related data, the Methods section should briefly explain the term “risk–benefit ratio” has no literal meaning. In addition, there may be reason for the omission (for example, “the design did not several distinct benefits and harms. We advocate using “balance ofbenefits and harms” rather than “risk–benefit ratio.” include the collection of any information on harms”).
Authors should also be explicit about separately re- Safety: Substantive evidence of an absence of harm. The term is often porting anticipated and unexpected adverse events. Expec- misused when there is simply absence of evidence of harm.
tation may influence the incidence of reported or ascer- Serious adverse events: As defined by the International Conference on tained adverse events. Making participants aware in the Harmonisation of Technical Requirements for Registration of consent form of the possibility of a specific adverse event Pharmaceuticals for Human Use, document E2A (available atwww.ich.org/UrlGrpServer.jser?@_IDϭ276&@_TEMPLATEϭ254): “During (“priming”) may increase the reporting rate of the event clinical investigations, adverse events may occur which, if suspected to be (20). Another example of priming is the finding that the medicinal product-related (adverse drug reactions), might be significant rates of withdrawals due to adverse events and the rates of enough to lead to important changes in the way the medicinal product isdeveloped (e.g., change in dose, population, needed monitoring consent specific adverse events were significantly higher in trials of forms). This is particularly true for reactions which, in their most severe aspirin, diclofenac, or indomethacin with comparator forms, threaten life or function. Such reactions should be reportedpromptly to regulators.” drugs compared with placebo-controlled trials (21). Pre-sumably, participants were more eager to come forth and Side effects: Unintended drug effects. The term, however, does not report an adverse event or to withdraw from treatment necessarily imply harm, as some side effects may be beneficial.
Furthermore, it tends to understate the importance of harms because when they knew they could not be receiving inactive pla- “side” may be perceived as denoting secondary importance.
Authors should report whether they used standardized Toxicity: Describes drug-related harms. The term may be most appropriate for laboratory-determined measurements, although it is also used in and validated measurement instruments for adverse events.
relation to clinical events. Abnormal laboratory values may be described Several medical fields have developed standardized scales as laboratory-determined toxicity. The disadvantage of the term“toxicity” is that it implies causality. If authors cannot prove causality, the (22–32). Use of nonvalidated scales is common. The terms “abnormal laboratory measurements” or “laboratory abnormalities” source document for well-established definitions and scales should be referenced. New definitions for adverse eventsshould be explicit and clear. Authors should describe howthey developed and validated new scales.
Authors should present information on harms in the For interventions that target healthy individuals (for abstract. If no important harms occurred, authors should example, many preventive interventions), any harm, how- so state. Explicit reference to the reporting of adverse ever minor, may be important to capture and report be- events in the title or abstract is also important for appro- cause the balance of harms and benefits may easily lean priate database indexing and information retrieval (19).
toward harms in a low-risk population. For other popula- 782 16 November 2004 Annals of Internal Medicine Volume 141 • Number 10
www.annals.org
Improving the Reporting of Harms Improving Patient Care Table 1. Original CONSORT Checklist
Paper Section and Topic
Descriptor
Reported on
Page Number
How participants were allocated to interventions (e.g., “random allocation”, “randomized”, or “randomly assigned”).
Scientific background and explanation of rationale.
Eligibility criteria for participants and the settings and locations where the data were Precise details of the interventions intended for each group and how and when they Clearly defined primary and secondary outcome measures and, when applicable, any methods used to enhance the quality of measurements (e.g., multipleobservations, training of assessors).
How sample size was determined and, when applicable, explanation of any interim Method used to generate the random allocation sequence, including details of any restriction (e.g., blocking, stratification).
Method used to implement the random allocation sequence (e.g., numbered containers or central telephone), clarifying whether the sequence was concealeduntil interventions were assigned.
Who generated the allocation sequence, who enrolled participants, and who assigned participants to their groups.
Whether or not participants, those administering the interventions, and those assessing the outcomes were blinded to group assignment. If done, how thesuccess of blinding was evaluated.
Statistical methods used to compare groups for primary outcome(s); Methods for additional analyses, such as subgroup analyses and adjusted analyses.
Flow of participants through each stage (a diagram is strongly recommended).
Specifically, for each group report the numbers of participants randomly assigned,receiving intended treatment, completing the study protocol, and analyzed for theprimary outcome. Describe protocol deviations from study as planned, togetherwith reasons.
Dates defining the periods of recruitment and follow-up.
Baseline demographic and clinical characteristics of each group.
Number of participants (denominator) in each group included in each analysis and whether the analysis was by “intention-to-treat”. State the results in absolutenumbers when feasible (e.g., 10/20, not 50%).
For each primary and secondary outcome, a summary of results for each group, and the estimated effect size and its precision (e.g., 95% confidence interval).
Address multiplicity by reporting any other analyses performed, including subgroup analyses and adjusted analyses, indicating those pre-specified and thoseexploratory.
All important adverse events or side effects in each intervention group.
Interpretation of the results, taking into account study hypotheses, sources of potential bias or imprecision and the dangers associated with multiplicity ofanalyses and outcomes.
Generalizability (external validity) of the trial findings.
General interpretation of the results in the context of current evidence.
tions and for interventions that improve major outcomes It is important to describe the questionnaires, inter- (for example, survival), severe and life-threatening adverse views, and tests used to collect information on harms, as events may be the only ones that are important in the well as their timing during follow-up. Passive surveillance of harms leads to fewer recorded adverse events than active Recommendation 4. Clarify how harms-related informa- surveillance (4). Open-ended questions may yield different tion was collected (mode of data collection, timing, attribution information, both quantitatively and qualitatively, than methods, intensity of ascertainment, and harms-related moni- structured questionnaires (33). Studies of nonsteroidal, toring and stopping rules, if pertinent). anti-inflammatory drugs (NSAIDs) exemplify how data www.annals.org
16 November 2004 Annals of Internal Medicine Volume 141 • Number 10 783
Improving Patient Care Improving the Reporting of Harms Table 2. Checklist of Items To Include When Reporting Harms in Randomized, Controlled Trials*
Standard CONSORT Checklist:
Standard CONSORT
Descriptor
Reported on
Paper Section and Topic
Checklist: Item Number
Page Number
If the study collected data on harms and benefits, the title or abstract should so state.
If the trial addresses both harms and benefits, the List addressed adverse events with definitions for each (with attention, when relevant, to grading,expected vs. unexpected events, reference tostandardized and validated definitions, anddescription of new definitions).
Clarify how harms-related information was collected (mode of data collection, timing, attributionmethods, intensity of ascertainment, andharms-related monitoring and stopping rules, ifpertinent).
Describe plans for presenting and analyzing information on harms (including coding, handlingof recurrent events, specification of timing issues,handling of continuous measures, and any statisticalanalyses).
Describe for each arm the participant withdrawals that are due to harms and their experiences withthe allocated treatment.
Provide the denominators for analyses on harms.
Present the absolute risk per arm and per adverse event type, grade, and seriousness, and present appropriate metrics for recurrent events, continuousvariables, and scale variables, whenever pertinent.† Describe any subgroup analyses and exploratory Provide a balanced discussion of benefits and harms with emphasis on study limitations, generalizability, and other sources of information on harms.‡ * This proposed extension for harms includes 10 recommendations that correspond to the original CONSORT checklist.
† Descriptors refer to items 17, 18, and 19.
‡ Descriptor refers to items 20, 21, and 22.
collection methods can affect the detection and reporting Authors should specify the time frame of surveillance of harms. When selective NSAIDs with fewer gastrointes- for adverse events. Some investigators stop recording ad- tinal adverse events became available, trials reported more verse events at the end of the intervention period or a than 10 times as many ulcers when comparing these drugs certain number of days afterward (for example, 30 days with older NSAIDs as when older NSAIDs were compared after discontinuation of drug therapy) and miss events with with placebo. In the newer trials, more ulcers were detected long latency (35). Surgical trials often capture only the because participants had regular endoscopy, and the case adverse events that occur intraoperatively. Several impor- definition of ulcers was more sensitive (34).
tant surgical complications are likely to occur later. Finally, 784 16 November 2004 Annals of Internal Medicine Volume 141 • Number 10
www.annals.org
Improving the Reporting of Harms Improving Patient Care in crossover trials, delayed events might occur while the longitudinal follow-up, specifying the timing of the events patient is taking a subsequent assigned treatment.
may be important (for example, to separate early from late Attribution is the process of deciding whether an ad- toxicity). Incidence rates, period prevalence rates, and point verse event is related to the intervention. Whenever authors prevalence rates may provide complementary information filter events through an attribution process, they should about the occurrence of an adverse event. Kaplan–Meier state who makes the attribution (investigators, participants, curves showing cumulative incidence of important adverse sponsors, or combinations), whether the process is blinded events can be helpful. Simple summaries with person-time to assigned treatment, and what definitions of adverse denominators (for example, median months after treat- ment) can be misleading if the event occurs only after Discontinuations and withdrawals due to adverse extended treatment and long follow-up, and most partici- events are especially important because they reflect the ul- pants had short follow-up and therefore no events.
timate decision of the participant and/or physician to dis- For continuous variables (such as reported for most continue treatment. Although treatment may occasionally laboratory tests), means and SDs or medians and inter- be discontinued for mild or moderate adverse events, at- quartile ranges may provide an aggregate picture, but they tributing discontinuation to a specific reason (to toxicity, may not convey information on extreme values that corre- lack of efficacy, other reasons, or combinations of reasons) spond to severe toxicity. Means and medians may be useful may be difficult. For example, in psychopharmacology, in informing participants and clinicians about expected, dropouts may reflect treatment ineffectiveness as much as toxicity-related intolerance (36). Trial reports should spec- Scales are increasingly used for measuring quality of ify who gave the reasons for discontinuation (participants life in RCTs. These measures are composite outcomes that or physicians) and whether attribution was blinded to the reflect both benefits and harms (39). Authors should de- assigned treatment. For example, even in blinded trials, scribe the development of these instruments, their validity participants and their clinicians are often unblinded before and sensitivity to detect change, and whether they assumed they decide whether to discontinue the intervention. It is interval scaling in order to use the scale as a continuous important to report participants who are nonadherent or lost to follow-up because their actions may reflect their inabil- When harms are major primary or secondary out- ity to tolerate the intervention. Moreover, authors should comes of a trial, the authors should describe plans to per- specify how they handled withdrawals in the analyses of the form any formal statistical analyses and inferences. They should separate prespecified statistical analyses from post Randomized, controlled trials should report any plan hoc analyses (40) and address common problems: low for monitoring for harms and rules for stopping the trial power for uncommon events, adjustment for multiple out- because of harms (37). They should clarify whether stop- comes, composite outcomes, regression to the mean (for ping guidelines examine benefits and harms separately or example, for laboratory tests that are also used for screening evaluate a composite measure that reflects the trade-off be- for study eligibility), and heterogeneity of treatment effects Statistical Methods
Participant Flow
Recommendation 5. Describe plans for presenting and Recommendation 6. Describe for each arm the partici- analyzing information on harms (including coding, handling pant withdrawals that are due to harms and the experience of recurrent events, specification of timing issues, handling of continuous measures, and any statistical analyses). Authors should describe the reasons for discontinua- Using only descriptive statistics to report harms is per- tions and reductions in dosage of the allocated treatment fectly appropriate in most RCTs because most trials lack and withdrawals from the study. They should emphasize power to test harms-related hypotheses and indeed have no harms-related reasons and acknowledge the caveats noted explicit prespecified harms-related hypotheses. If investiga- under Recommendation 6. Authors should always report tors combine data for different adverse events into 1 out- deaths in each study group during a trial, regardless of come measure, they should describe each combination, cite whether death is an end point and regardless of whether the dictionary that lists the definitions of the adverse attribution to a specific cause is possible (42).
events, and state whether they decided the grouping of Randomized, controlled trials with prolonged fol- low-up should report the timing of allocated treatment The distributions of adverse events over the follow-up received, dose reductions and discontinuations, and study period can pose problems for analysis of the data. When withdrawals. The cause of early withdrawals may differ pertinent, authors should specify whether they count recur- from that of late withdrawals; separate descriptions of each rent events (events that occur more than once in the same may enhance the accuracy of information on the tolerabil- participant) as separate events or as 1 event. For trials with ity profile of an intervention. Kaplan–Meier plots of the www.annals.org
16 November 2004 Annals of Internal Medicine Volume 141 • Number 10 785
Improving Patient Care Improving the Reporting of Harms Table 3. Common Poor Reporting Practices for
Moreover, authors should state whether they use the same Harms-Related Data
type of analysis for both efficacy end points and harms.
1. Using generic or vague statements, such as “the drug was generally well tolerated” or “the comparator drug was relatively poorly Rates of Outcomes and Ancillary Analyses for Adverse Events
Recommendation 8. Present the absolute risk of each ad- 2. Failing to provide separate data for each study arm.
verse event (specifying type, grade, and seriousness per arm),and present appropriate metrics for recurrent events, continu- 3. Providing summed numbers for all adverse events for each study arm, ous variables and scale variables, whenever pertinent. without separate data for each type of adverse event.
Authors should present results separately for each 4. Providing summed numbers for a specific type of adverse event, study group of a trial. For each type of adverse event, they regardless of severity or seriousness.
should offer appropriate metrics of absolute risk (for exam- 5. Reporting only the adverse events observed at a certain frequency or ple, frequency or incidence), with separate information rate threshold (for example, Ͼ3% or Ͼ10% of participants).
about the severity grade of the event, if relevant. Seriousevents should be reported separately for each type of event.
6. Reporting only the adverse events that reach a P value threshold in the comparison of the randomized arms (for example, P Ͻ 0.05).
Recurrent events and timing of events need appropriatereporting, as discussed in Recommendation 5. For events 7. Reporting measures of central tendency (for example, means or with many recurrences, it is useful to provide both the medians) for continuous variables without any information on extremevalues.
number of affected participants and the number of eventsfor each study group and rate (events per unit of person- 8. Improperly handling or disregarding the relative timing of the events, time). Occasionally, a graphical representation of the dis- when timing is an important determinant of the adverse event inquestion.
tribution of number of events per patient or time-to-eventanalyses may be informative.
9. Not distinguishing between patients with 1 adverse event and Overall, the Results section should report on what the participants with multiple adverse events.
Methods section promises (43). Any break from this sym- 10. Providing statements about whether data were statistically significant metry requires explanation. If no adverse events of a spe- without giving the exact counts of events.
cific type and severity occurred, authors should so state in
the Results section (44). Table 3 shows common reporting
11. Not providing data on harms for all randomly assigned participants.
Recommendation 9. Describe any subgroup analyses and time-to-discontinuation or time-to-withdrawal for each study group may be useful, especially if the treatment du- Reporting of adverse events for different participant ration is considerably different from the time of follow-up.
subgroups follows the same principles that govern the re- These plots might reflect (and graphically show) differ- porting of subgroup analyses for efficacy. Authors should ences in adverse events between treatment groups, which avoid overstating the significance of false-positive subgroup could suggest that harm is associated with the intervention.
findings (45). Authors should state how, why, and whenthey planned subgroup analyses (a priori or post hoc). Reg-ulatory agencies increasingly require subgroup analyses by Numbers Analyzed
age, sex, and race for license applications. However, these Recommendation 7. Provide the denominators for analy- variables rarely show any significant effect modification for efficacy outcomes (45, 46) and may be equally low-yield There are many ways to report the number of partic- ipants included in analyzing harms. Randomized, con-trolled trials in which time-on-treatment differs from total Discussion
follow-up should report the denominator for each analysis Recommendation 10. Provide a balanced discussion of (that is, which participants and what follow-up time count benefits and harms with emphasis on study limitations, gener- toward total exposure to the allocated treatment. Termi- alizability, and other sources of information on harms. nology such as “intention-to-treat,” “modified intention- The Discussion is typically the most poorly structured to-treat,” “available-case,” and “on-treatment” analyses, section of an RCT report and could be modified to gain can be confusing and may provide different results, de- space for reporting harms. Authors may curtail the length pending on the type of analyses conducted. Overall, inten- of the Discussion section to gain space for appropriate re- tion-to-treat is usually the preferred analysis both for effi- cacy and harms because intention-to-treat is an analysis in In summarizing the key findings of an RCT, the Dis- which the original random participant assignment is main- cussion section should pay attention to the original trial tained in the data analysis. Because differences in the use of objectives and provide a balanced view that puts the ben- the definitions of these types of analysis can be important, efits and harms into perspective. Authors should avoid authors should state which analyses and definitions they use.
overinterpretation of the findings. Limitations are probably 786 16 November 2004 Annals of Internal Medicine Volume 141 • Number 10
www.annals.org
Improving the Reporting of Harms Improving Patient Care the most important part of the Discussion section. Com- readers to submit comments, critique, and suggestions for mon limitations in studies that report harms include in- improvement through www.consort-statement.org. We conclusive findings, lack of power, multiplicity of compar- also hope that journals and editorial groups will support isons, post hoc analyses, and short duration of exposure to our efforts to improve the reporting of harms. We ask the allocated treatment, especially for treatments of chronic journals that endorse the CONSORT reporting require- diseases. Generalizability is often a problem for harms. The ment to include a reference to this document in their In- frequency and severity of adverse events may depend on structions to Authors section. Adherence to reporting stan- the clinical setting and participants. Often, clinical trials dards for harms should help to inform readers and the enroll participants who have the disease of interest but are public on the harms of interventions.
otherwise healthy and do not have comorbid conditions.
Once licensed, however, most approved interventions are From University of Ioannina School of Medicine and Biomedical Re- used in individuals who have several comorbid conditions search Institute, Foundation for Research and Technology–Hellas, Ioan- and who are taking several other drugs with potentially nina, Greece; London School of Hygiene and Tropical Medicine, Lon-don, United Kingdom; The Nordic Cochrane Centre, Copenhagen, Denmark; U.S. Food and Drug Administration, Rockville, Maryland; The Discussion section should also appraise emerging Cancer Research UK/National Health Service Centre for Statistics in data on benefits and harms. Authors should systematically Medicine, Oxford, United Kingdom; Family Health International, Re- integrate prior evidence on harms, whenever possible (47).
search Triangle Park, North Carolina; and Children’s Hospital of East- If a systematic review of previous studies of harms is not ern Ontario Research Institute, Ottawa, Ontario, Canada.
possible, authors should so state, perhaps to stimulatesomeone to correct the deficiency in the future. Authors Disclaimer: Dr. O’Neill is an employee of the U.S. Food and Drug
should contrast the trial results on harms with other Administration. The views expressed are those of the author and do notnecessarily represent those of the U.S. Food and Drug Administration.
sources of information on harms, including observationaldata from spontaneous reporting, automated databases, Grant Support: Abbott Laboratories, Children’s Hospital of Eastern On-
case– control studies, and case reports.
tario Research Institute, Health Canada, The Lancet, and Merck and Co., Manuscript Length
Inc. provided funding for the Montebello meeting. Dr. Altman is sup-ported by Cancer Research UK.
Improved reporting of harms need not lead to longer manuscripts. In the Methods section, adopting standard Potential Financial Conflicts of Interest: None disclosed.
definitions with appropriate references may actually savespace. Tables may summarize key results on harms. Graphs Requests for Single Reprints: David Moher, PhD, Chalmers Research
may convey important time-to-event outcomes or repeated Group, Children’s Hospital of Eastern Ontario Research Institute, 401Smyth Road, Ottawa, Ontario K1H 8L1, Canada; e-mail, dmoher measurements of adverse events. Finally, it is possible to write short Discussion sections by using an appropriatestructure (48).
Current author addresses are available at www.annals.org.
Occasionally, investigators present adverse events sepa- rately in another paper. This practice denies both author and References
reader the opportunity to formulate the balance between ben- 1. Cuervo LG, Clarke M. Balancing benefits and harms in health care [Editorial].
efits and harms. Therefore, authors should report harms and benefits together in the same manuscript. Authors and journal 2. Altman DG, Schulz KF, Moher D, Egger M, Davidoff F, Elbourne D, et al.
editors should publish additional trial information on the The revised CONSORT statement for reporting randomized trials: explanation Web, as an adjunct to the main results. For example, Web and elaboration. Ann Intern Med. 2001;134:663-94. [PMID: 11304107]
3. Moher D, Schulz KF, Altman D. The CONSORT statement: revised recom-
material could include enhanced graphical displays that mendations for improving the quality of reports of parallel-group randomized present individual participants’ experiences (49).
trials. JAMA. 2001;285:1987-91. [PMID: 11308435] Investigators may sparingly use single patient reports 4. Stephens MD, Talbot JC, Routledge PA, eds. The Detection of New
embedded in the results of an RCT to describe severe, Adverse Reactions. 4th ed. London: Macmillan Reference; 1998.
serious, and previously unreported adverse events. More 5. Edwards IR, Aronson JK. Adverse drug reactions: definitions, diagnosis, and
management. Lancet. 2000;356:1255-9. [PMID: 11072960]
comprehensive case reports may require a separate paper.
6. Ioannidis JP, Contopoulos-Ioannidis DG. Reporting of safety data from ran-
In the RCT report, a single sentence may suffice to de- domised trials [Letter]. Lancet. 1998;352:1752-3. [PMID: 9848355] scribe an unusual adverse event by succinctly summarizing 7. Ioannidis JP, Lau J. Completeness of safety reporting in randomized trials: an
the type of adverse event, when it occurred, the type of evaluation of 7 medical areas. JAMA. 2001;285:437-43. [PMID: 11242428] patient, the management of the adverse event, and the out- 8. Gøtzsche PC. Methodology and overt and hidden bias in reports of 196
double-blind trials of nonsteroidal antiinflammatory drugs in rheumatoid arthri-
tis. Control Clin Trials. 1989;10:31-56. [PMID: 2702836]
9. Loke YK, Derry S. Reporting of adverse drug reactions in randomised con-
CONCLUSIONS
trolled trials—a systematic survey. BMC Clin Pharmacol. 2001;1:3. [PMID:11591227] This extension of the main CONSORT statement to 10. Edwards JE, McQuay HJ, Moore RA, Collins SL. Reporting of adverse
include harms is a work in progress. We therefore invite effects in clinical trials should be improved: lessons from acute postoperative pain.
www.annals.org
16 November 2004 Annals of Internal Medicine Volume 141 • Number 10 787
Improving Patient Care Improving the Reporting of Harms J Pain Symptom Manage. 1999;18:427-37. [PMID: 10641469] Pharmacoepidemiol Drug Saf. 2003;12:335-40. [PMID: 12812014] 11. Hayashi K, Walker AM. Japanese and American reports of randomized trials:
29. Division of AIDS. Table for Grading Severity of Adult Adverse Experiences.
differences in the reporting of adverse effects. Control Clin Trials. 1996;17:99- Rockville, MD: National Institute of Allergy and Infectious Diseases; 1992.
30. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of
12. Papanikolaou PN, Churchill R, Wahlbeck K, Ioannidis JP. Safety reporting
cancer treatment. Cancer. 1981;47:207-14. [PMID: 7459811] in randomized trials of mental health interventions. Am J Psychiatry. 2004;161: 31. Common Toxicity Criteria, version 2. Bethesda, MD: National Cancer In- 13. Martin RC 2nd, Brennan MF, Jaques DP. Quality of complication report-
32. Corso DM, Pucino F, DeLeo JM, Calis KA, Gallelli JF. Development of a
ing in the surgical literature. Ann Surg. 2002;235:803-13. [PMID: 12035036] questionnaire for detecting potential adverse drug reactions. Ann Pharmacother.
14. Ioannidis JP, Chew P, Lau J. Standardized retrieval of side effects data for
meta-analysis of safety outcomes. A feasibility study in acute sinusitis. J Clin 33. Wallin J, Sjovall J. Detection of adverse drug reactions in a clinical trial using
Epidemiol. 2002;55:619-26. [PMID: 12063104] two types of questioning. Clin Ther. 1981;3:450-2. [PMID: 7020939] 15. Ioannidis JP, Lau J. Improving safety reporting from randomised trials. Drug
34. Gøtzsche PC. Non-steroidal anti-inflammatory drugs. BMJ. 2000;320:1058-
16. Mucklow JC. Reporting drug safety in clinical trials: getting the emphasis
right. Lancet. 2001;357:1384. [PMID: 11356432]
35. Roehrborn CG. Reporting of acute urinary retention in BPH treatment
trials: importance of patient follow-up after discontinuation and case definitions.
17. Lesko SM, Mitchell AA. An assessment of the safety of pediatric ibuprofen.
A practitioner-based randomized clinical trial. JAMA. 1995;273:929-33. [PMID:7884951] 36. Wahlbeck K, Tuunainen A, Ahokas A, Leucht S. Dropout rates in random-
ised antipsychotic drug trials. Psychopharmacology (Berl). 2001;155:230-3.
18. Castle W, Fuller R, Hall J, Palmer J. Serevent nationwide surveillance study:
comparison of salmeterol with salbutamol in asthmatic patients who require reg-ular bronchodilator treatment. BMJ. 1993;306:1034-7. [PMID: 8098238] 37. Whitehead J. On being the statistician on a Data and Safety Monitoring
19. Derry S, Kong Loke Y, Aronson JK. Incomplete evidence: the inadequacy of
Board. Stat Med. 1999;18:3425-34. [PMID: 10611616] databases in tracing published adverse drug reactions in clinical trials. BMC Med 38. Thall PF, Cheng SC. Optimal two-stage designs for clinical trials based on
safety and efficacy. Stat Med. 2001;20:1023-32. [PMID: 11276033] 20. Myers MG, Cairns JA, Singer J. The consent form as a possible cause of side
39. Sanders C, Egger M, Donovan J, Tallon D, Frankel S. Reporting on quality
effects. Clin Pharmacol Ther. 1987;42:250-3. [PMID: 3621780] of life in randomised controlled trials: bibliographic study. BMJ. 1998;317: 21. Rochon PA, Binns MA, Litner JA, Litner GM, Fischbach MS, Eisenberg
D, et al. Are randomized control trial outcomes influenced by the inclusion of a
40. Smith A. Report of the Statisticians in the Pharmaceutical Industry Working
placebo group? A systematic review of nonsteroidal antiinflammatory drug trials for arthritis treatment. J Clin Epidemiol. 1999;52:113-22. [PMID: 10201651] 41. Chuang-Stein C. Laboratory data in clinical trials: a statistician’s perspective.
22. Peloso PM, Wright JG, Bombardier C. A critical appraisal of toxicity in-
Control Clin Trials. 1998;19:167-77. [PMID: 9551281] dexes in rheumatology. J Rheumatol. 1995;22:989-94. [PMID: 8587096] 42. Cato A. Premarketing adverse drug experiences: data management proce-
23. Marshall M, Lockwood A, Bradley C, Adams C, Joy C, Fenton M. Un-
dures. Unexpected death occurring early in clinical trials. Drug Inf J. 1987;21: published rating scales: a major source of bias in randomised controlled trials of reatments for schizophrenia. Br J Psychiatry. 2000;176:249-52. [PMID: 43. Chan AW, Hro´bjartsson A, Haahr MT, Gøtzsche PC, Altman DG. Em-
pirical evidence for selective reporting of outcomes in randomized trials: compar- 24. Vitiello B, Riddle MA, Greenhill LL, March JS, Levine J, Schachar RJ, et al.
ison of protocols to published articles. JAMA. 2004;291:2457-65. [PMID: How can we improve the assessment of safety in child and adolescent psychop- harmacology? J Am Acad Child Adolesc Psychiatry. 2003;42:634-41. [PMID: 44. Jovanovic BD, Zalenski RJ. Safety evaluation and confidence intervals when
the number of observed events is small or zero. Ann Emerg Med. 1997;30:301-6.
25. Bonhoeffer J, Kohl K, Chen R, Duclos P, Heijbel H, Heininger U, et al.
The Brighton Collaboration: addressing the need for standardized case definitions 45. Oxman AD, Guyatt GH. A consumer’s guide to subgroup analyses. Ann
of adverse events following immunization (AEFI). Vaccine. 2002;21:298-302.
Intern Med. 1992;116:78-84. [PMID: 1530753] [PMID: 12450705]
26. Woodworth TG, Furst DE, Strand V, Kempeni J, Fenner H, Lau CS, et al.
46. Yusuf S, Wittes J, Probstfield J, Tyroler HA. Analysis and interpretation of
Standardizing assessment of adverse effects in rheumatology clinical trials. Status treatment effects in subgroups of patients in randomized clinical trials. JAMA.
of OMERACT Toxicity Working Group March 2000: towards a common un- derstanding of comparative toxicity/safety profiles for antirheumatic therapies.
47. Clarke M, Alderson P, Chalmers I. Discussion sections in reports of con-
J Rheumatol. 2001;28:1163-9. [PMID: 11361207] trolled trials published in general medical journals. JAMA. 2002;287:2799-801.
27. Trotti A, Bentzen SM. The need for adverse effects reporting standards in
oncology clinical trials [Editorial]. J Clin Oncol. 2004;22:19-22. [PMID: 48. Docherty M, Smith R. The case for structuring the discussion of scientific
papers [Editorial]. BMJ. 1999;318:1224-5. [PMID: 10231230] 28. Kohl KS, Bonhoeffer J, Chen R, Duclos P, Heijbel H, Heininger U, et al.
49. O’Neill RT. Regulatory perspectives on data monitoring. Stat Med. 2002;
The Brighton Collaboration: enhancing comparability of vaccine safety data.
788 16 November 2004 Annals of Internal Medicine Volume 141 • Number 10
www.annals.org
APPENDIX 1: MEMBERS OF THE CONSORT GROUP
Title of trial with primary harms outcome: Kamran Abbasi, MB (British Medical Journal); Douglas Alt- man, DSc (Centre for Statistics in Medicine, Oxford, United An assessment of the safety of pediatric ibuprofen. A Kingdom); Matthias Egger, MD (Bristol University, Bristol, practitioner-based randomized clinical trial (17).
United Kingdom, and University of Bern, Bern, Switzerland;International Journal of Epidemiology); Diana Elbourne, PhD (London School of Hygiene and Tropical Medicine, London, No benefit, but increased harm from high dose (100 United Kingdom); Stephen J.W. Evans, MSc (London School of ␮g) misoprostol for induction of labour: a randomised Hygiene and Tropical Medicine, London, United Kingdom); trial of high vs. low (50 ␮g) dose misoprostol (51).
Joel Gagnier, ND (University of Toronto, Toronto, Ontario,Canada); Peter Gøtzsche, MD (Nordic Cochrane Centre, Copenhagen, Denmark); David Grimes, MD (Family Health In-ternational, Research Triangle Park, North Carolina); John Ioan- There were two uterine ruptures and four intrapartum nidis, MD (University of Ioannina, Ioannina, Greece); David stillbirths in the high misoprostol group. There was no Jaques, MD (Memorial Sloan-Kettering Cancer Center, New difference in postpartum haemorrhage, 9.5% vs. 7.9% York, New York); Tom Lang, MA (Tom Lang Communications, (P ϭ 1.00) and admissions to the neonatal unit 18.8% Lakewood, Ohio), Joseph Lau, MD (Tufts-New England Medi- vs. 17.0% (P ϭ 0.980) in the 1ow- and high-groups) cal Center, Boston, Massachusetts); Yoon Kong Loke, MD (Uni- respectively. . . . the higher dose had an increased risk versity of Oxford, Oxford, United Kingdom); Faith McLellan, PhD (The Lancet); Siddika Mithani, MD (Health Canada–Clini-cal Trials, Ottawa, Ontario, Canada); David Moher, PhD (Chalmers Research Group and University of Ottawa, Ottawa,Ontario, Canada); Cynthia Mulrow, MD (Annals of Internal CONCLUSION: In postmenopausal women with cor- Medicine and University of Texas Health Sciences Center, San onary disease, neither hormonal replacement therapy Antonio, Texas); Robert O’Neill, PhD (Center for Drug Evalu- nor antioxidant vitamin supplements provide cardio- ation and Research, U.S. Food and Drug Administration, Rock- vascular benefit. Instead, a potential for harm was sug- ville, MD); Gilda Piaggio, PhD (World Health Organization, Geneva, Switzerland); Drummond Rennie, MD (University of Recommendation 2
California at San Francisco, San Francisco, California; Journal ofthe American Medical Association); David Riley, MD (University Short courses of prednisone, 1 to 2 mg/kg daily for 3 to of New Mexico Medical School, Santa Fe, New Mexico; EX- 10 days, are recommended for the management of PLORE: The Journal of Science and Healing); David Sackett, MD acute exacerbations of asthma in children. . . . We (Trout Research and Education Centre, Irish Lake, Ontario, could find no evidence in the literature that 2 mg/kg is Canada); Roberta Scherer, PhD (University of Maryland School more beneficial than 1 mg/kg in treating children. Fur- of Medicine, Baltimore, Maryland); Kenneth Schulz, PhD (Fam- thermore, studies in asthmatic children have looked ily Health International, Chapel Hill, North Carolina); James primarily at efficacy and parental management prac- Scott, MD (University of Utah, Salt Lake City, Utah; Obstetrics tices, rather than at the possible adverse effects. Because and Gynecology). Observers: Don Husereau, BScPharm, MSc
the parents of many of our patients have voiced con- (Canadian Coordinating Office for Health Technology, Ottawa, cern about the behavioral changes in their children dur- Ontario, Canada); Andra Morrison, BSc (Chalmers Research ing short courses of treatment with oral steroids, we Group, Ottawa, Ontario, Canada); Isabelle French, BSc (Dal- decided to conduct a prospective, randomized, blinded housie University, Halifax, Nova Scotia, Canada); Gabriela Lewin, trial comparing the adverse effects and the benefits atthe two dose levels (53).
MD (Chalmers Research Group, Ottawa, Ontario, Canada).
Recommendation 3
Note: Laurence Hirsch, MD, and Beate Stych, MD (Merck), partici-
. . . assessments included minimum systolic blood pres-sure, minimum pressure of oxygen, maximum concen- APPENDIX 2: EXAMPLES OF RECOMMENDATIONS
tration of delivered nitrogen dioxide, and maximum Recommendation 1
concentration of methemoglobin (54).
Comprehensive list with definitions: The safety and efficacy of prophylactic ondansetron in Bleeding complications and other adverse events were patients undergoing modified radical mastectomy (50).
documented by interview or were reported by the pa- www.annals.org
Annals of Internal Medicine Volume 141 • Number 10 W-147
tients during the study. Major bleeding was defined as level lower boundaries for adverse effects. The adverse- any clinically apparent bleeding associated with a de- effect boundaries were further adjusted with a Bonfer- crease of at least 2.0 g per dL in the hemoglobin level, roni correction for the 7 major outcomes other than requirement for transfusion of at least 2 units of packed breast cancer that were specifically monitored (58).
red cells, or retroperitoneal or intracranial bleeding or Recommendation 5
other bleeding that the investigators decided requiredpermanent discontinuation of treatment. Bleeding that did not meet this definition was considered minor. Anadverse event was considered serious if it was fatal or More than 200 distinct types of clinical, laboratory or life threatening, caused permanent disability, or required ECG events were noted using . . . literal description, hospitalization or prolonged hospitalisation (55).
but the FDA coding symbols for the Thesaurus of Ad-verse Reaction Terms (COSTART), used to report ad- Definitions and grading (referral to established system): verse events, reduced that number to 110. Since someadverse events were similar, i.e. fatigue, asthenia, feeling . . . using the AIDS Clinical Trials Group adverse event unease, they were also regrouped, resulting in . . . 37 Timing issues— early vs. late events: In June 1998, the protocol was amended to include Early reactions. Early reactions were recorded by the laboratory monitoring, toxicity management, and dose nursing staff in the cardiac catheter laboratory and on reduction of adefovir dipivoxil for proximal renal tubu- the cardiology ward after the patient left the catheter lar dysfunction (PRTD), because of new information laboratory. . . . Late reactions. Each patient was asked to about the toxicity provided by Gilead Sciences. The complete a simple questionnaire after discharge from protocol definition of PRTD was as follows: serum cre- the hospital, on which to record any adverse reactions atinine 0.5 mg/dL above baseline and serum phos- occurring within 1 week of the cardiac catheterization.
phate Ͻ 2.0 mg/dL, or 1 of these abnormalities plus 2 . . . The analysis of patients with late skin reactions was of the following: proteinuria (2ϩ), glycosuria (1ϩ) in confined to patients with reactions that had clearly the absence of hyperglycemia, hypokalemia (Ͻ3.0 started after hospital discharge and therefore were not a mEq/L), or serum bicarbonate Ͻ19 mEq/L (57).
continuation of, for example, an urticarial reaction oc- Recommendation 4
curring in the catheter laboratory. . . . (61) [Note: use ofthe term “reactions” may not be optimal, since it implies causality, but the example is appropriate for timing issues].
At each semiannual contact, a standardized interview Continuous measures (mean estimates and serious extremes): collected information on designated symptoms and[harms] concerns, and initial reports of outcome events . . . assessed by changes in vital signs (summarized as a were obtained using a self-administered questionnaire mean [SE] change from baseline) and by reports of adverse events with onset within 8 weeks of random-ization. All reports of adverse events were included whether or not they were deemed by the investigator tobe related to treatment. An adverse event was defined in Adverse experiences and toxic effects were assessed ev- the study protocol as serious if it was fatal or life-threat- ery 4 weeks until 12 weeks after the discontinuation of ening, required or prolonged hospitalization, or result- ed in persistent or significant disability or incapacity (62).
Causality was assessed by the investigator at the time of Patients reported occurrence and severity of 18 side the event, using a modified version of Karch and Lasa- effects. . . . Between-group differences were compared Recommendation 6
Trial monitoring guidelines for early stopping consid- Withdrawals and harm-related reasons for withdrawals per arm: erations were based on O’Brien-Fleming boundaries us-ing asymmetric upper and lower boundaries: a 1-sided, Eight other patients (6 treated with fluconazole and 2 .025-level upper boundary for benefit and 1-sided, .05- treated with itraconazole) were withdrawn from the W-148 16 November 2004
Annals of Internal Medicine Volume 141 • Number 10 www.annals.org
study because of mild to moderate symptoms, such as Appendix Table 1. Adverse Events among Human
rash (fluconazole group), dry skin (fluconazole group), Immunodeficiency Virus-Infected Patients Whose Aphthous
nausea (fluconazole group), or difficulty concentrating Ulcers Had Healed Previously When Treated with Thalidomide
and Who Then Were Treated with Thalidomide or Placebo in a
Maintenance-Phase Study
Adverse Event
Patients Taking
Patients Taking
No patients receiving trimethoprimϪsulfamethoxazole Thalidomide
(n ؍ 23),
(n ؍ 26),
or ciprofloxacin discontinued therapy with the drug be- Grade 3 or higher
A substantial number of women had stopped taking study drugs at some time (42% of estrogen plus pro- gestin and 38% of placebo) [Figure shows cumulative dropout rates over time] (58).
Any grade
The study drug exposure was as follows: 15% received 1 dose; 31% received 2 doses; 37% received 3 doses; Recommendation 7
* Data represent patients with at least 1 episode of an adverse event during studytreatment (68). Adapted with permission from Jacobson JM, Greenspan JS, Spritz- All randomly assigned and treated patients (n ϭ 68) ler J, Fox L, Fahey JL, Jackson JB, et al. Thalidomide in low intermittent dosesdoes not prevent recurrence of human immunodeficiency virus–associated aph- were included in the . . . [harms] . . . analysis (67).
thous ulcers. J Infect Dis 2001;183:343–346.
Recommendation 8
P ϭ 0.002) [table also shown]. On average, each patient Absolute risks for binary events per arm and per type and grade experienced 2.8 (95% CI, 1.9 to 3.7) fewer episodes of (follow-up/exposure time approximately comparable for all partici- hypoglycemia with split than with mixed dosing [Fig- See Appendix Table 1.
Continuous measures presented with information on both aver- Absolute risks for binary events per arm and per type and grade (follow-up/exposure time is differential and not comparable for all There were small changes in mean (SE) systolic blood pressure (placebo group, Ϫ2.0 [0.5] mm Hg and carve- 45 patients (46 events) in the rofecoxib group and 20 Appendix Table 2. Adverse Events during the First 8 Weeks*
patients (20 events) in the naproxen group were adju- Adverse Event
Patients Taking
Patients Taking
dicated to have serious thrombotic cardiovascular ad- Carvedilol
verse events (myocardial infarction, unstable angina, (n ؍ 1133),
(n ؍ 1156),
cardiac thrombus, resuscitated cardiac arrest, sudden or unexplained death, ischemic stroke, and transient isch- Bradycardia
emic attacks). Event-free survival analysis of these 66 patients showed that the RR (95% confidence interval [CI]) of developing a cardiovascular event in the rofe- coxib treatment group was 2.38 (1.39-4.00), P Ͻ .001 [Figure shows the Kaplan–Meier plots for time to cardio- Hypotension
vascular adverse event in each arm] (69).
Recurrent events expressed with person-time denominator: The frequency of hypoglycemia at 3:00 a.m. wasgreater in the mixed-treatment period than in the split- * Table is presented in part. Adapted with permission from Krum H, Roecker EB, treatment period (0.28 [SD 0.04] episode/patient-day Mohacsi P, Rouleau JL, Tendera M, Coats AJ, et al. Effects of initiating carvedilolin patients with severe chronic heart failure; results from the COPERNICUS vs. 0.10 [SD 0.02] episode/patient-day, respectively; www.annals.org
Annals of Internal Medicine Volume 141 • Number 10 W-149
dilol group, Ϫ3.6 [0.5] mm Hg) and in diastolic blood Tropical Medicine, Keppel Street, London WC1E 7HT, United King- pressure (placebo group, Ϫ1.8 [0.3] mm Hg and carve- dilol group, Ϫ2.7 [0.3] mm Hg) at the end of 8 weeks.
Dr. Gøtzsche: The Nordic Cochrane Centre, Rigshospitalet, Depart- . . .Patients in the carvedilol group were more likely ment 7112, Blegdamsvej 9, DK-2100 Copenhagen, Denmark.
than in the placebo group to report . . . hypotension Dr. O’Neill: Center for Drug Evaluation and Research, U.S. Food andDrug Administration, 5600 Fishers Lane, Rockville, MD 20857.
. . . and bradycardia [see Appendix Table 2] (62).
Dr. Altman: Cancer Research UK/NHS Centre for Statistics in Medi- Recommendation 9
cine, Old Road Campus, Old Road, Headington, Oxford OX3 7LF,United Kingdom.
Dr. Schultz: Family Health International, PO Box 13950, Research Tri- Across the same subgroups shown in Figures 1–3, the relative risk of “any bleeding” with treatment compared Dr. Moher: Chalmers Research Group, Children’s Hospital of Eastern with placebo ranged from 0.65 to 1.86 (data not Ontario Research Institute, 401 Smyth Road, Ottawa, Ontario K1H shown). No statistically significant treatment-by-sub- group interactions were noted . . . Using the same con-sistency criteria employed for mortality, “any bleeding” Web-Only References
and “serious bleeding” results for all subgroups wereconsistent with the overall trial results. For bleeding 50. Sadhasivam S, Saxena A, Kathirvel S, Kannan TR, Trikha A, Mohan V.
The safety and efficacy of prophylactic ondansetron in patients undergoing mod-
event end points (any or serious bleeding), there was no ified radical mastectomy. Anesth Analg. 1999;89:1340-5. [PMID: 10589605] statistically significant interaction with predicted risk of 51. Majoko F, Nystrom L, Lindmark G. No benefit, but increased harm from
mortality (P ϭ .55 and P ϭ .21) (72).
high dose (100 microg) misoprostol for induction of labour: a randomised trial ofhigh vs. low (50 microg) dose misoprostol. J Obstet Gynaecol. 2002;22:614-7.
Recommendation 10
[PMID: 12554247]
52. Waters DD, Alderman EL, Hsia J, Howard BV, Cobb FR, Rogers WJ, et
al. Effects of hormone replacement therapy and antioxidant vitamin supplements
on coronary atherosclerosis in postmenopausal women: a randomized controlled
trial. JAMA. 2002;288:2432-40. [PMID: 12435256]
[In] our study pretreatment with a small dose of sub- 53. Kayani S, Shannon DC. Adverse behavioral effects of treatment for acute
cutaneous adrenaline significantly reduced the inci- exacerbation of asthma in children: a comparison of two doses of oral steroids.
dence of acute adverse reactions to polyvalent anti- venom serum. . . . We did not encounter significant 54. Weiner DL, Hibberd PL, Betit P, Cooper AB, Botelho CA, Brugnara C.
adverse effects attributable to it; there were no cases of Preliminary assessment of inhaled nitric oxide for acute vaso-occlusive crisis inpediatric patients with sickle cell disease. JAMA. 2003;289:1136-42. [PMID: acute neurological deficit suggestive of cerebrovascular accidents or patients in whom blood pressure rose sig- 55. Lassen MR, Borris LC, Nakov RL. Use of the low-molecular-weight heparin
reviparin to prevent deep-vein thrombosis after leg injury requiring immobiliza-tion. N Engl J Med. 2002;347:726-30. [PMID: 12213943] 56. Hammer SM, Vaida F, Bennett KK, Holohan MK, Sheiner L, Eron JJ, et
al. Dual vs single protease inhibitor therapy following antiretroviral treatment
failure: a randomized trial. JAMA. 2002;288:169-80. [PMID: 12095381]
Even this large study was still too small to allow for a 57. Gulick RM, Hu XJ, Fiscus SA, Fletcher CV, Haubrich R, Cheng H, et al.
rigorous . . . assessment [of harms]. The data on drug Randomized study of saquinavir with ritonavir or nelfinavir together with dela- related clinical adverse experiences show that etoricoxib virdine, adefovir, or both in human immunodeficiency virus-infected adults with may be better tolerated than indometacin [sic], but ad- virologic failure on indinavir: AIDS Clinical Trials Group Study 359. J Infect ditional studies are needed to enable any definitive con- 58. Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, Kooperberg C,
Stefanick ML, et al. Risks and benefits of estrogen plus progestin in healthy
postmenopausal women: principal results from the Women’s Health Initiative
randomized controlled trial. JAMA. 2002;288:321-33. [PMID: 12117397]
59. Englund JA, Baker CJ, Raskino C, McKinney RE, Petrie B, Fowler MG, et
Adefovir dipivoxil-related PRTD occurred in 1% of al. Zidovudine, didanosine, or both as the initial treatment for symptomatic
study patients during the first 16 weeks. In other stud- HIV-infected children. AIDS Clinical Trials Group (ACTG) Study 152 Team.
N Engl J Med. 1997;336:1704-12. [PMID: 9182213] ies, 35% of patients taking adefovir dipivoxil developed 60. Sibille M, Deigat N, Janin A, Kirkesseli S, Durand DV. Adverse events in
significant increases in serum creatinine levels, and phase-I studies: a report in 1015 healthy volunteers. Eur J Clin Pharmacol. 1998; 50% developed significant hypophosphatemia, labora- tory abnormalities consistent with PRTD, by 48 weeks 61. Sutton AG, Finn P, Grech ED, Hall JA, Stewart MJ, Davies A, et al. Early
and late reactions after the use of iopamidol 340, ioxaglate 320, and iodixanol320 in cardiac catheterization. Am Heart J. 2001;141:677-83. [PMID:11275937] Current Author Addresses: Dr. Ioannidis: Clinical Trials and Evidence-
62. Krum H, Roecker EB, Mohacsi P, Rouleau JL, Tendera M, Coats AJ, et al.
Based Medicine Unit, Department of Hygiene and Epidemiology, Uni- Effects of initiating carvedilol in patients with severe chronic heart failure: results versity of Ioannina School of Medicine, and Biomedical Research Insti- from the COPERNICUS Study. JAMA. 2003;289:712-8. [PMID: 12585949] tute, Foundation for Research and Technology-Hellas, Ioannina 45110, 63. Pandya KJ, Raubertas RF, Flynn PJ, Hynes HE, Rosenbluth RJ, Kirshner
JJ, et al. Oral clonidine in postmenopausal patients with breast cancer experienc-
Mr. Evans: Medical Statistics Unit, London School of Hygiene and ing tamoxifen-induced hot flashes: a University of Rochester Cancer Center W-150 16 November 2004
Annals of Internal Medicine Volume 141 • Number 10 www.annals.org
Community Clinical Oncology Program study. Ann Intern Med. 2000;132:788- 69. Mukherjee D, Nissen SE, Topol EJ. Risk of cardiovascular events associated
with selective COX-2 inhibitors. JAMA. 2001;286:954-9. [PMID: 11509060] 64. Galgiani JN, Catanzaro A, Cloud GA, Johnson RH, Williams PL, Mirels
70. Fanelli CG, Pampanelli S, Porcellati F, Rossetti P, Brunetti P, Bolli GB.
LF, et al. Comparison of oral fluconazole and itraconazole for progressive, non-
Administration of neutral protamine Hagedorn insulin at bedtime versus with meningeal coccidioidomycosis. A randomized, double-blind trial. Mycoses Study dinner in type 1 diabetes mellitus to avoid nocturnal hypoglycemia and improve Group. Ann Intern Med. 2000;133:676-86. [PMID: 11074900] control. A randomized, controlled trial. Ann Intern Med. 2002;136:504-14.
65. Verdier RI, Fitzgerald DW, Johnson WD Jr, Pape JW. Trimethoprim-
sulfamethoxazole compared with ciprofloxacin for treatment and prophylaxis of 71. Krum H, Roecker EB, Mohacsi P, Rouleau JL, Tendera M, Coats AJ, et al.
Isospora belli and Cyclospora cayetanensis infection in HIV-infected patients. A ran- Effects of initiating carvedilol in patients with severe chronic heart failure: results domized, controlled trial. Ann Intern Med. 2000;132:885-8. [PMID: 10836915] from the COPERNICUS Study. JAMA. 2003;289:712-8. [PMID: 12585949] 66. Corwin HL, Gettinger A, Pearl RG, Fink MP, Levy MM, Shapiro MJ, et
72. Ely EW, Laterre PF, Angus DC, Helterbrand JD, Levy H, Dhainaut JF, et
al. Efficacy of recombinant human erythropoietin in critically ill patients: a ran-
al. Drotrecogin alfa (activated) administration across clinically important sub-
domized controlled trial. JAMA. 2002;288:2827-35. [PMID: 12472324] groups of patients with severe sepsis. Crit Care Med. 2003;31:12-9. [PMID: 67. Seibold JR, Korn JH, Simms R, Clements PJ, Moreland LW, Mayes MD,
et al. Recombinant human relaxin in the treatment of scleroderma. A random-
73. Premawardhena AP, de Silva CE, Fonseka MM, Gunatilake SB, de Silva
ized, double-blind, placebo-controlled trial. Ann Intern Med. 2000;132:871-9.
HJ. Low dose subcutaneous adrenaline to prevent acute adverse reactions to
antivenom serum in people bitten by snakes: randomised, placebo controlled 68. Jacobson JM, Greenspan JS, Spritzler J, Fox L, Fahey JL, Jackson JB, et al.
trial. BMJ. 1999;318:1041-3. [PMID: 10205101] Thalidomide in low intermittent doses does not prevent recurrence of human 74. Schumacher HR Jr, Boice JA, Daikh DI, Mukhopadhyay S, Malmstrom K,
immunodeficiency virus-associated aphthous ulcers. J Infect Dis. 2001;183:343- Ng J, et al. Randomised double blind trial of etoricoxib and indometacin in
treatment of acute gouty arthritis. BMJ. 2002;324:1488-92. [PMID: 12077033] www.annals.org
Annals of Internal Medicine Volume 141 • Number 10 W-151

Source: http://consort-uat.koneka.com/Media/Default/Downloads/Extensions/CONSORT%20Extension%20for%20Harms.pdf