A prospective, randomized, double-blind controlled trial of acetaminophen and diphenhydramine for the prevention of transfusion reactions

BackgroundStudy done in North Carolina with

bedside leukoreductionRates of febrile reactions:

RBC transfusions: 0.3-6% With prestorage leukoreduction: 0.2%

Platelet transfusions: 1-38% With prestorage leukoreduction: 0.2%

BackgroundCanada has prestorage leukoreductionRates of febrile reactions:

RBC transfusions: 0.3% Platelet transfusions: 10%

O’Brien et al. Current incidence and estimated residual risk of transfusion-transmitted infections in donations made to Canadian Blood Services. Transfusion 2007;47:316-325.

Pathophysiology The recipient’s leukocyte antibodies form

antibody-antigen complexes with the donor leukocytes

This interaction actives effector cells (monocytes and B cells) to produce and release IL-1, IL-6, TNF-α and other proinflammatory mediators

These same inflammatory substances can accumulate in blood products during storage and cause febrile reactions in the absence of recipient WBC antibodies

Background Isolated chills and rigors without fever

were not considered febrile reactions in this study

BackgroundStudy

Rates of allergic transfusion reactions: 0.4-3% Not mitigated by leukoreduction

Canada Rates of allergic transfusion reactions:

1%

Pathophysiology Recipient antibodies react with plasma

proteins or other substances in the donor unit Preformed recipient IgE on mast cells and

basophils interacting with this antigen leads to activation and degranulation Mast cell and basophil degraulation releases

histamine, adenosine, chemotactic factors and enzymes resulting in allergic symptoms

QuestionMost febrile reactions respond to

acetaminophen (and allergic reactions to diphenhydramine)

Can we use these medications prophylactically to prevent these reactions?

Why is this important?These reactions can be uncomfortable

or distressing for the patientThese reactions are very common and

have financial impact Utilizes limited nursing resources Utilizes physician resources May increase product utilization

Early severe reactions may be confused with these common, minor reactions

What are the disadvantages? 68-80% of patietns will be prophylaxed Acetaminophen may mask fever unrelated to

transfusion (infection) Rare side effects include hepatotoxicity

Sedation with diphenhydramine may be bothersome to an otherwise active patient

Cost ($40,000) Ethical considerations

MethodRandomized, double-blind placebo

controlled study315 BMT patients aged 18-65Exclusions:

Allergy to the study medications Documented history of febrile or allergic

transfusion reaction

MethodRandomized by the pharmacist to

receive either 500 mg acetaminophen and 25 mg diphenhydramine or placebo 30 minutes before RBC or platelet transfusion Using blocked randomization

All caregivers/members of the team were blinded throughout the study duration and data collection

MethodThe PI reviewed the medical record to

determine if a transfusion occurredPatients were removed from the study

once a transfusion reaction was documented

Statistical analysisThe study was designed to accrue 320

patients 90% power for detecting a HR of 0.4

(treatment relative to placebo) at the 10% one-sided level of significance

Assumes 10% of the patients would have reactions

Equal type 1 and type 2 error rates were chosen because it was equally important to protect against falsely rejecting or falsely accepting the null hypothesis

Results

ResultsThere was no difference between the

number of reactions in the placebo versus the experimental arm (p=0.433)

ProblemsThe rate of reactions are not in keeping

with expected rates FNHTRs: 0.62%

Non-traditional definition of FNHTRs Allergic: 0.86%

Problems “A log rank test was used to assess the

unadjusted difference between groups in the number of transfusions received before a reaction was noted” “If the Kaplan-Meier survival curves cross

then this is clear departure from proportional hazards, and the log rank test should not be used”

BMJ Statistics at Square One http://www.bmj.com/collections/statsbk/12.dtl

ProblemsTherefore, we can’t rely on the

conclusion: “the number of transfusions until a febrile reaction was significantly greater (at the pre-specified 0.1 level of significance) for patients receiving the active drug (one sided p=0.074)” Invalid conclusion: “the standard practice

. . . may reduce febrile reactions”

ProblemsNo transfusions were excluded when

study drugs were administered off-study 33% of transfusions were administered

under these circumstances Re-analysis excluding these transfusions

and results were similar except for the fact that the rate of reactions was higher

Insufficient power Does this violate intention to treat?

What were the results?Prophylactic administration of

acetaminophen and diphenhydramine does not significantly decrease the rate of transfusion reactions in BMT patients receiving RBCs or platelets who have not previously had an allergic reaction or FNHTR

How large was the treatment effect?There was no difference between the

frequency of allergic reactions or FNHTRs in the experimental arm versus the placebo arm The authors claim the number of febrile

transfusion reactions received before the first febrile reaction was significantly greater for patients receiving the active drug than for placebo

Are the results valid?Have the results been systematically

biased? Excluded patients with previous reactions Patients removed from study after first

reaction Average of 13 transfusions/patient Removal of “reactors”

Was the assignment of patients to treatment randomized?

Patients were randomized by the pharmacist using blocked randomization

What is randomization? The process of assigning participants to

treatment groups in a known but unpredictable fashion The participant should have an equal chance of

being assigned to any of the treatment groups It helps ensure that the treatment and control

groups will have similar characteristics of both known and unknown factors Any difference between groups will occur only by

chance (avoids systematic bias)

Different types of randomizationFixed randomization

Probability of allocation to each treatment group remains constant

Adaptive randomization Probability of being assigned to a

treatment group changes as a function of such variables as the number of patients assigned to the group, the subjects’ baseline characteristics, or observed outcomes

Fixed allocation randomizationAssigns the intervention to participants

with a pre-specified probability, usually equal, and that allocation probability is not altered as the study progresses

Simple randomizationBlocked randomizationStratified randomization

Simple randomizationSimple randomization

Coin toss or random number generator Advantages include simplicity and in the

long run probability dictates that the groups will be equivalent

Disadvantages are that for small to medium sample sizes, can end up with very unequal results

If randomizing 100 patients, there in only an 8% chance that there will be 50 patients in either group

Blocked Randomization In a trial of 60 subjects there could be 10

consecutive blocks of 6, each containing 3 allocations to the control group and 3 allocations to the treatment group

Advantages: ensures the same number of patients in both the control and experimental groups Even if the study is stopped early, the maximum

imbalance in sample size is half the size of the block

Disadvantages: more complicated analysis

Blocked RandomizationThere are six different ways to allocate

four patients to two groups AABB ABAB ABBA BABA BAAB BBAA

Blocked randomizationRoll a dice to choose the allocation

pattern for the upcoming blockEach of the six patterns has the same

likelihood of being chosenThis pattern guarantees that the groups

will be balanced after every 4 patients The maximum imbalance between the

groups is 2 patients

Stratified randomization Stratified random allocation involves first

identifying important prognostic factors and then separately randomising blocks containing different levels of the prognostic factor

The prognostic factors that are most commonly stratified are disease severity and, in multi-site, trials, the site at which the subject is treated

Stratified randomization Advantages

Potentially increases statistical power

Disadvantages it is not practically possible to stratify by all

important prognostic factors stratification must be blocked if it is to be useful,

but it is difficult to match block sizes so that each stratum fills at approximately the same time

as the trial nears completion, the researchers may have to discontinue recruitment into one stratum while they wait for another stratum to fill up

Were all patients who entered the trial properly accounted for at its conclusion?All transfusion events were accounted

for Even those events which occurred while

study medications were being used off-study

No “loss to follow up” Events were analyzed in the groups to

which they were randomized

Were patients, their clinicians and study personnel “blind” to treatment?All study personnel (other than the

pharmacist) were blinded to the allocation of patients

Were the groups similar at the start of the trial?

Aside from the experimental intervention, were the groups treated equally?Presumably there should have been no

difference in the way that the two groups were managed

Will the results help me in caring for my patients?

Can the results be applied to my patient care?The patient population (BMT patients)

accounts for a substantial portion of patients receiving blood products in our hospital

The authors excluded patients who had had previous reactions, which is a population of particular interest

Were all clinically important outcomes considered?Used criteria (somewhat restrictive for

febrile) to determine whether patients did or did not have reactions

It might be interesting to also have had subjective patient comfort as an outcome

Are the likely treatment benefits worth the potential harm and cost? In this case, the authors were unable to

show any difference in the rates of transfusion reactions (allergic and FNHTRs) between the patients who received premedication and those who did not

Therefore, the treatment has no benefit Potential harm is likely to be minimal The impact on cost is probably moderate

Conclusion In conclusion, there is no evidence the

prophylaxis with acetaminophen and diphenhydramine prevents allergic or FNHTRs in BMT patients who have not experienced a previous reaction Recommendation: Do not prophylax BMT

patients with no history of reaction, but treat reactions as they occur