*Tenecteplase group that received alteplase placebo. †Alteplase
group that received tenecteplase placebo.
Adjusted for age, Killip class, age × Killip class,
systolic blood pressure, heart rate, anterior myocardial infarction, previous
myocardial infarction, height, weight, symptom duration, diabetes, current
smoking, prior coronary artery bypass graft surgery, and hypertension. *Tenecteplase
group that received alteplase placebo.
Mehta RH, Alexander JH, Van de Werf F, Armstrong PW, Pieper KS, Garg J, Califf RM, Granger CB. Relationship of Incorrect Dosing of Fibrinolytic Therapy and Clinical Outcomes. JAMA. 2005;293(14):1746–1750. doi:10.1001/jama.293.14.1746
Author Affiliations: Duke Clinical Research
Institute and Duke University Medical Center, Durham, NC (Drs Mehta, Alexander,
Califf, and Granger and Mss Pieper and Garg); Department of Internal Medicine,
Division of Cardiology, University Hospital Leuven, Leuven, Belgium (Dr Van
de Werf); and Department of Internal Medicine, Division of Cardiology, University
of Alberta, Edmonton (Dr Armstrong).
Context Incorrect dosing of alteplase has been associated with worse clinical
outcomes in patients. However, patients at high risk of adverse events are
more prone to dosing errors, thus confounding this relationship.
Objective To determine if the association between incorrect dosing of alteplase
and adverse outcomes is related to cause and effect or to confounding.
Design, Setting, and Patients Observational analysis in May 2004 of a double-blind, double-dummy trial
of 16 949 patients with ST-segment elevation myocardial infarction who
were enrolled in the Assessment of the Safety and Efficacy of a New Thrombolytic
(ASSENT-2) trial and were assigned to either a bolus of tenecteplase (with
alteplase placebo bolus plus infusion) or a bolus of alteplase (with tenecteplase
placebo plus infusion).
Main Outcome Measures Thirty-day mortality, in-hospital stroke, and major bleeding associated
with incorrect dosing of active alteplase compared with placebo alteplase.
Results Incorrect dosing occurred in 4.9% of patients who received active alteplase
and in 4.6% of patients who received alteplase placebo. Patients receiving
incorrect doses of alteplase or alteplase placebo were more likely to be older,
female, black, shorter, have lower body weight and systolic blood pressure,
and have a higher Killip class at presentation. Thirty-day mortality was higher
in patients who received an overdose (9.8%) or underdose (19.5%) of alteplase
compared with those who received a correct dose (5.4%). The same pattern was
present in patients who received an alteplase placebo (10.0% for overdose,
23.5% for underdose, and 5.4% for correct dose). Similar patterns were seen
for in-hospital intracranial hemorrhage and major bleeding. The higher rates
of adverse outcomes with incorrect dosing were largely accounted for by adjusting
for baseline characteristics.
Conclusions The relationship between incorrect dosing and patient outcome in ASSENT-2
is primarily due to confounding factors rather than incorrect dosing itself.
These data highlight the need for caution when ascribing a causal relationship
to associations between incorrect dosing and adverse outcomes.
Incorrect dosing of fibrinolytic therapy has been reported to occur
5% to 12% of the time.1- 3 Several
studies have reported higher mortality, stroke, and major hemorrhagic event
rates in patients who received incorrect doses of fibrinolytic agents.4- 7 However,
several patient factors identified as related to risk of incorrect dosing8 are also markers of higher risk of death, thereby
limiting inference about the cause and effect relationship of incorrect dosing
and adverse outcomes. It is assumed that the adverse outcomes are caused by
incorrect dosing. However, it is also possible that the adverse outcomes may
be due to confounding factors, such that sicker patients with an unstable
early clinical course could be more likely to receive incorrect doses.
To determine how much of the association between incorrect dosing and
adverse outcomes is cause and effect, we examined the Assessment of the Safety
and Efficacy of a New Thrombolytic (ASSENT-2) trial database.9 Our
hypothesis was that if the incorrect dose was causing adverse outcomes, the
association between incorrect dosing of active fibrinolytic and adverse outcome
would be much stronger than the association between incorrect dosing of fibrinolytic
placebo. Conversely, if incorrect dosing of a fibrinolytic placebo and adverse
outcome were equally related to incorrect dosing of an active fibrinolytic
drug, the association would be mostly due to confounding.
After providing informed consent, eligible patients in the ASSENT-2
trial (n = 16 949) were randomly assigned in a double-blind,
double-dummy fashion to receive (1) a body weight–adjusted bolus of
tenecteplase (5- to 10-second bolus; range 30-50 mg) plus a bolus and infusion
of placebo (alteplase placebo) or (2) a bolus and infusion of alteplase plus
a bolus of placebo (tenecteplase placebo).9 Alteplase
and alteplase placebo were given as 15-mg boluses and were followed by a 0.75-mg/kg
infusion (≤50 mg) over 30 minutes and a 0.50-mg/kg infusion (≤35 mg)
over 60 minutes. All patients received 150 to 325 mg of aspirin orally and
intravenous heparin adjusted to an activated partial thromboplastin time of
50 to 75 seconds for 48 to 72 hours. Incorrect dosing was defined as any dose
more or less than the weight-based dose of fibrinolytic agent.9 Each
participating center obtained approval from its local ethics board prior to
patient enrollment. The race of each patient was noted by a check mark on
the case report form selected by the investigator.
Baseline and end point variables are presented as frequencies and percentages
for categorical factors and as medians with 25th and 75th percentiles for
continuous variables. To evaluate the differences in outcomes between the
various comparison treatment groups, previously established multivariable
models to predict 30-day mortality,10 in-hospital
stroke,11 and in-hospital major bleeding12 in patients with ST-segment elevation myocardial
infarction (STEMI) were used to adjust for the influence of confounders when
calculating adjusted odds ratios (ORs). P<.01
was considered significant because of the large numbers of patients and multiple
comparisons. All statistical analyses were performed in May 2004 using SAS
software version 8.0 (SAS Institute Inc, Cary, NC).
Incorrect dosing occurred in 4.9% of patients receiving alteplase; 4.6%,
alteplase placebo (tenecteplase group); 3.6%, tenecteplase; and 3.6%, tenecteplase
placebo (alteplase group). Underdosing occurred in 3.2% of patients receiving
alteplase, 3.1% of patients receiving alteplase placebo, and 1.8% of patients
receiving tenecteplase and tenecteplase placebo. Overdosing occurred in 1.7%
of patients receiving alteplase, 1.5% of patients receiving alteplase placebo,
and 1.8% of patients receiving tenecteplase and tenecteplase placebo. The
median degree of underdosing was 15% (4%-25%) less than the correct dose and
the median degree of overdosing was 5% (2%-10%) more than the correct dose.
The incidence of incorrect dosing in the same patient for alteplase and tenecteplase
placebo occurred in 0.24%, while the incidence of incorrect dosing in the
same patient for alteplase placebo and active tenecteplase occurred in 0.19%.
Characteristics of patients and clinical events among those with and
without incorrect dosing of alteplase are shown in Table 1 and Table 2. Compared
with patients receiving the correct dose, those receiving either overdoses
or underdoses of alteplase were more likely to be older, female, black, shorter,
have lower body weight and systolic blood pressure, and have a higher Killip
class, which are high-risk characteristics.
Incorrect alteplase dosing was associated with an increased risk for
30-day mortality (Figure 1). However,
the risk was increased to a nearly identical extent with incorrect dosing
of the alteplase placebo. After adjustment for confounding baseline features,
overdosing was no longer significantly associated with higher mortality for
either alteplase or alteplase placebo groups compared with correct dosing
of alteplase or alteplase placebo (Figure 2).
Compared with patients who received the correct doses of alteplase and alteplase
placebo, the unadjusted mortality risk was significantly higher for patients
who received underdoses of alteplase (OR, 4.27; 95% confidence interval [CI],
3.10-5.89) or alteplase placebo (OR, 5.42; 95% CI, 3.98-7.37). However, the
OR decreased substantially with adjustment for both underdosing of alteplase
and alteplase placebo. When deaths on the first day were excluded to remove
any effect from rapidly fatal complications, which may have caused the study
drug to be discontinued early, the impact of underdosing on 30-day mortality
was no longer statistically significant in either groups receiving underdoses
of alteplase or alteplase placebo compared with the respective correct dosing
groups. Furthermore, no difference in death was observed between the underdose
alteplase group and underdose alteplase placebo group or between the overdose
alteplase group and overdose alteplase placebo group (Figure 1 and Figure 2).
Similar to that observed for 30-day mortality, the same pattern of increased
risk with incorrect dosing of alteplase and of alteplase placebo was seen
for major bleeding and intracranial hemorrhage (Table 2). The excess risk of stroke and major bleeding with incorrect
dosing in the alteplase group and the alteplase placebo group also appeared
to be explained by higher baseline risk profile and was attenuated significantly
after adjustments for confounders (Table 3).
Finally, the relationship between incorrect dosing of tenecteplase placebo
and 30-day mortality demonstrated directional consistency similar to that
seen with incorrect dosing of alteplase placebo (Figure 1).
Our study shows that even in clinical trials where protocols are specified
and a select group of STEMI patients are enrolled, incorrect dosing of alteplase
occurred in approximately 5% of patients.1,3,13 While
we observed nearly the same excess risk of death, intracranial hemorrhage,
and bleeding with incorrect dosing of alteplase that has been reported in
prior studies,4- 7 this
excess risk was nearly identical with both incorrect dosing of alteplase placebo
and active alteplase. This suggests that most, if not all, of the association
of incorrect dosing of alteplase and adverse outcome was due to confounding
rather than directly resulting from incorrect dosing. In fact, when we adjusted
for these confounding factors, the risks of these clinical events were significantly
attenuated. Importantly, we identified several patient-related factors associated
with incorrect dosing of alteplase including advanced age, female sex, black
race, shorter stature, lower body weight, lower systolic blood pressure, and
a higher Killip class. These factors also have been shown to be associated
with increased risk of adverse clinical events in STEMI patients.10- 12 This may be related
to the fact that an unstable acutely ill patient with STEMI, heart failure,
or impending shock, and other comorbid conditions may require attention to
various aspects of his/her care simultaneously for stabilization, which distracts
physicians from reperfusion treatment itself and leads to dosing errors.13 The finding of more bleeding with underdosing highlights
the limitation of observational analysis. However, this finding may have resulted
from early discontinuation of fibrinolytic therapy in patients with early
The complexity of dosing and impact of early clinical instability were
reduced as potential confounders with dosing of tenecteplase and tenecteplase
placebo given as a 10-second bolus. Thus, the proportion of error as well
as the relationship of incorrect dosing and adverse outcome with this simple
single bolus administration was lower.
Medical errors due to incorrect dosing of fibrinolytic therapy have
been shown to be associated with increased risk of adverse events in STEMI
patients and with increased risk of litigation against caregivers.10- 13 When
identifying an incorrect dose of a potentially toxic drug associated with
an adverse outcome, the reflex and logical conclusion is to assume cause,
particularly in malpractice litigation that is principally driven by adverse
outcomes. However, this study raises the possibility that much of the adverse
outcomes ascribed to dosing errors, at least in some situations, may be due
to confounding rather than a direct effect of the error itself.
While education, appropriate use of technology, expertise, and a systematic
approach provide the best potential opportunity for reducing dosing errors,
physicians should also be encouraged to pay particular attention to avoid
incorrect dosing for patients with the characteristics that are predictive
of errors being made.
Inference regarding a lack in the cause and effect relationship between
incorrect dosing and adverse outcomes should not be considered definitive
because of the observational nature of our analysis. The relationship between
underdosing of alteplase and alteplase placebo may be confounded by the fact
that drug may be discontinued early because of an adverse event. Nevertheless,
our data provide reasonable evidence that adverse events in patients treated
with incorrect doses of fibrinolytic agents are not always solely related
to incorrect dosing, but they may also be attributable to patient factors
that are otherwise associated with poor outcomes. The clinical trial setting
differs from general practice and thus our results may not be generalizable
to clinical care in which one might expect that error rates would, if anything,
be higher. Most of the errors we describe are of modest proportions, and larger
errors would likely cause harm. However, these errors would likely also be
subject to considerable confounding.
Our data indicate that incorrect dosing of alteplase for patients with
STEMI was associated with much worse outcomes. The finding that the same relationship
between worse outcome and incorrect dose was seen with placebo, and that it
was attenuated when adjusted for baseline risk features, suggests that the
adverse outcomes are due to confounding rather than cause and effect. Thus,
while medication errors need to be minimized, caution should be used in concluding
that adverse outcomes associated with errors are primarily caused by the errors.
Importantly, more research is needed into dosing errors, with the same rigor
as in other clinical research, to better understand factors related to such
errors and their impact on patient outcome.
Corresponding Author: Christopher B. Granger,
MD, Box 3409, Duke University Medical Center, Durham, NC 27710 (firstname.lastname@example.org).
Author Contributions: Dr Granger had full access
to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analysis.
Study concept and design: Mehta, Alexander,
Acquisition of data: Alexander, Armstrong,
Analysis and interpretation of data: Mehta,
Alexander, Van de Werf, Armstrong, Pieper, Garg, Granger.
Drafting of the manuscript: Mehta, Armstrong,
Critical revision of the manuscript for important
intellectual content: Mehta, Alexander, Van de Werf, Armstrong, Pieper,
Garg, Califf, Granger.
Statistical expertise: Mehta, Pieper, Garg,
Obtained funding: Armstrong, Granger.
Administrative, technical, or material support:
Alexander, Armstrong, Califf, Granger.
Study supervision: Van de Werf, Armstrong,
Financial Disclosures: Dr Alexander has received
research funding from Boehringer Ingelheim. Dr Van de Werf has served on the
speaker’s bureau and has received research grants from Boehringer Ingelheim
and Genentech Inc. Drs Armstrong and Granger have received research funding
from Boehringer Ingelheim and Genentech Inc. None of the other authors reported
Funding/Support: The ASSENT-2 trial was funded
by Boehringer Ingelheim and Genentech Inc, but the sponsors did not provide
specific funding for this analysis.
Role of the Sponsors: Boehringer Ingelheim
and Genentech Inc had no role in the design or conduct of the study, in the
collection, analysis, interpretation of the data, or in the preparation, review,
or approval of the manuscript.