Download Original submitted manuscript 990542 – compare to final published

Original submitted manuscript 990542 – compare to final published
version and reviews
Copyright American College of Emergency Physicians
Improved outcomes in patients with acute allergic syndromes who are treated with
combined H1 and H2 antagonist
Abstract
Study objective: Although the addition of H2 blockers to H1 antagonists have been
promoted for use in anaphylaxis, there have been no large studies establishing the
advantage of this approach in treating acute allergic syndromes. In this study we tested
the hypothesis that combined H1 and H2 blockage results in improved outcomes in
patients treated for acute allergic syndromes compared to treatment with H1 blockade
alone.
Methods: In a randomized, double-blind, placebo controlled trial, 91 adult patients with
acute allergic syndromes were treated with either diphenhydramine (50 mg) and saline
(control group) or with diphenhydramine (50 mg) and ranitidine (50 mg)(active group).
These patients were treated by parenteral administration. Patients were recruited from an
emergency department at an urban academic medical center. The primary endpoints were
resolution of urticaria, angioedema, or erythema at 2 hours time after protocol treatment.
Areas of cutaneous involvement, heart rates, blood pressure, respiratory findings, and
symptom scores were also assessed at baseline, I hours, and 2 hours.
1
Results: There were significantly more patients without urticaria at 2 hours time in
patients in the active group compared to the control group. Both groups had similar
proportions of urticaria at baseline. Logistic regression models to predict resolution of
urticaria which accounted for baseline urticarial involvement showed odds ratios in favor
of the active group treatment. Similar findings were observed when the absence of both
urticaria and angioedema was considered as the dependent variable. There was not a
significant difference between the 2 groups with regards to the absence of erythema or
angioedema (irrespective of urticaria presence) at 2 hours. Blood pressure, and symptoms
did not show differences between the 2 groups over time. Heart rates decreased during
the first hour to a greater extent in the active treatment group compared to the placebo
group. Consistent with this observation, multivariate analysis showed that, after adjusting
for significant covariates, there was a significant effect of treatment over time on heart
rates.
Conclusion: These data show that adding H2 blockers to H1 antagonists results in
additional improvement of certain clinical outcomes for patients presenting with acute
allergic syndromes. This findings favor the recommendation for using combined H1 and
H2 anti-histamines in acute allergic syndromes.
2
Introduction:
Although pretreatment with combined H1 and H2 blockade has been shown to
have advantages over H1 blockade alone in preventing allergic reactions (1-5) and in
diminishing certain manifestations associated with parenteral histamine infusion(6), there
have been only one small study (7) examining outcomes in treating patients with ongoing
allergic reactions. Indeed, since allergic reactions may be associated with the elevation of
mediators other than histamine, it is conceivable that merely blocking more histamine
receptors with H2 antagonists would not result in clinical improvement due to a relatively
smaller contribution of histamine to the manifestations in these patients. Mast cell or
basophil release of histamine may result from specific IgE directed towards allergens.
Alternatively non-IgE stimuli can stimulate histamine release resulting in identical
clinical manifestations. Because mast cell or basophil release cannot be readily
determined in the acute setting, most allergic reactions presenting to the emergency
department are diagnosed on physical signs and symptoms in the setting of a consistent
history. Histamine release can be varied in terms of the presenting physical signs.
Cutaneous signs such as urticaria and angioedema predominate. However, flushing alone
or the complete absence of cutaneous signs may occur (5). Given the heterogeneity and
of these signs (some of which are highly nonspecific for histamine release), it is likely
that allergic syndromes diagnosed on clinical grounds may misclassify patients in terms
of actual histamine release associated manifestations. Anti-histamine administration has
constituted one of the mainstays in treating acute allergic syndromes, and due to the
relative safety of H1 antagonists, this approach to non-critically ill patients is reasonable.
Patients who may not have histamine release may still benefit symptomatically from anti-
3
histaminic therapy. Some patients have more allergic symptoms, such as those with
anaphylaxis, or they may have recurrence of presenting manifestations after initial H1
antagonist therapy. In these patients, the question of additional anti-allergic treatments is
especially pertinent. The safety profile of H2 antagonists is excellent, and it is reasonable
to consider the use of this agent not only in more severe cases but also for the purposes of
expediting clinical improvement in the emergency department setting. However the
additional benefit of H2 antagonists when given with H1 antagonists has never been
examined in emergency settings in a large head to head controlled study. For this reason,
we decided to determine whether combined H1 and H2 blockade treatment results in
improved outcomes in patients presenting with acute allergic syndromes. We used a
rather broad definition of allergic syndromes in order to approximate real life emergency
department approaches to patients with various symptoms and signs.
4
Methods:
Adult patients (>18 years old) were considered for recruitment from the
Emergency Department if they the following syndromes following an ingested food or
ingested, inhaled, or injected drug or after contact with latex acute urticaria, acute
angioedema, acute unexplained stridor, and acute pruritic rash. These manifestations
should have been present for no greater than 12 hours from the time of alleged allergen
exposure. Pregnant patients were excluded.
Recruited patients were randomized to treatment with either diphenhvdramine 50
mg and ranitidine 50 mg (active treatment group) or diphenhydramine 50 mg and normal
saline(control treatment group), in a convenience sample based on study physician
scheduling(representing approximately 1/3 of all available attending physician time
slots). Each treatment designation was placed in sealed, opaque envelopes stored in a
locked cabinet. A staff member uninvolved with the patient’s care then drew into a
syringe either 50 mg of raniditine solution or an equal volume of normal saline based on
a random number assignment list. Immediately after an intravenous injection of 50 mg
diphenhydramine was given, the unmarked syringe containing either saline or ranitidine
was then given to the study physician, unaware of its contents. The physician then
administered the contents by intravenous injection to the subject. Supplemental
medications such as epinephrine, corticosteroids, bronchodilators, and additional doses of
antihistamines as well as intravenous fluids were administered at the discretion of the
study physicians. Patients also received supplemental oxygen and intravenous fluids at
the discretion of the study physicians. Patients had heart rate, blood pressure, physical
5
findings, side effects, and symptoms assessed at baseline, 1 hour, and 2 hours relative to
experimental treatment. Baseline temperatures were also recorded. Physical findings that
were recorded included the presence and extent of urticaria and erthema, and the presence
of angioedema, wheezing, stridor, abdominal distention or tenderness, and abdominal
hyperactive bowel sounds. Historical features, physical findings (including heart rates,
blood pressure and respiratory rates), and treatments were recorded on a study specific
data input form. The extent of involvement with urticaria and erythema was assessed
using a check off cartoon of body areas (similar to that used to assess burn area extent)
printed on the data input sheet. Symptom scores were assessed at baseline, 1 hour, and 2
hours, using a preprinted form with none, mild, moderate, and severe check off
categories.
Since no prior studies existed to determine an estimate of effect size (in terms of
resolution of urticaria and angioedema), an arbitrary sample size of 100 was used The
primary variables of interest were resolution of urticaria, and angioedema at 2 hours time.
Changes in heart rates, respiratory rates, blood pressure, and symptoms were also
examined. The final disposition of the patient was also noted (admission, discharge, or
leaving against medical advice). The study was approved by the Institutional Review
Board and informed written consent was obtained from all patients.
The resolution of urticaria, erythema, and angioedema was assessed using
bivariate chi squared analysis and by multivariate logistic regression. Potential
confounding factors and pertinent covariates were included in some multivariate models.
6
The changes in heart rates, blood pressure, respiratory rates, and symptoms were
analyzed by repeated measured analysis of variance or covariance (8), with the time by
group interaction being the analysis of primary interest (9). The Greenhouse-Geisser
adjustment for non-sphericity was applied (10). Heart rates were normalized by
logarithmic transformation for analysis. Analyses were performed using the JMPTM
software package (Version 3.2)(SAS Institute, Cary NC). Certain statistical values are
expressed with 95% confidence intervals.
7
Results:
Among 100 patients recruited for the study, one was inadvertently studied twice.
Only the first study for this patient was used. Eight other patients withdrew after the
study medication had been obtained but before administration was given. Nine patients
received experimental treatment but upon chart review had allergic manifestations for
greater than 12 hours (2 patients each at 15 hours and 48 hours, 1 each at 19, 40, 58, 69,
and 120 hours) . These patients were included in the analysis on an intention-to-treat
principle. The patients recruited over a 12 month period between May 1998 and April
1999. There was no predominance of weekend or particular shift recruitment.
The baseline characteristics of these patients are shown in part in Table I. No
significant baseline differences were observed between the 2 groups. No patients had
fever (T>38.8 C). Two patients had pretreatment blood pressures less than 90/60 (both in
the raniditine group). Respiratory and gastrointestinal signs were present in many patients
(Tables I and III). Wheezing in 12 patients was observed in the absence of a history of
asthma. Non-atopic comorbid conditions were reported by 6 and 10 patients in the
placebo and ranitidine groups respectively. These included hypertension (3),
hypercholesterolemia (3), HIV infection (2), diabetes mellitus (2), thyroid disorders (2),
substance abuse (1), cardiac disease(3), gastroesophageal reflux (l), seizures(1), and
psychiatric disorders (2). Three patients were taking oral birth control pills. Ten patients
had taken prior antihistamines within 36 hours of the study (4 in the raniditine group and
6 in the placebo group). There was no predominance of any particular comorbiditv in
8
either of the 2 treatment groups. There was a high incidence of reported asthma and nonasthmatic atopic conditions in both groups (Table I).
More than half of the patients (n=58) had a suspected food allergic reaction. Of
these patients, shellfish was implicated in 14, and fish, fruit, and vegetables were
implicated in 9 patients each. Nut or peanut allergic reactions were suspected in 5
patients. Allergic reactions due to drug administration was suspected in 28 patients. Of
these, anti-infectives were suspected in 11, non-steroidal anti-inflammatory drugs in 8,
and psychiatric/neurologic drugs in 4. The causative agent in 5 patients was uncertain as
to whether food or medication related to the allergic syndrome.
Additional treatments were given to the study patients as detailed in Table II.
There was no difference in the proportion of patients who received intravenous fluids
between the 2 groups (data not shown). There were significantly more patients in the
placebo group who received extra antihistamines. Seven patients were admitted to the
hospital (3-ranitidine, 4-placebo). No adverse effects from the protocol treatment were
observed.
The proportions of patients with urticaria at baseline were similar between the 2
groups. For those patients with baseline urticaria, the numbers of areas involved with
urticaria was not significantly different between the 2 groups. The number of patients
with urticaria decreased somewhat more in the ranitidine group at 1 hour. The proportion
of patients without urticaria was significantly greater in the ranitidine group (91.7%) at 2
hours compared to the placebo group (73.8%)(p =0.2). Furthermore the number of areas
9
involved with urticaria at 2 hours was significantly less in the raniditine group compared
to the placebo group (Wilcoxon sign rank test, p= 0.2). In multivariate modeling, the
number of areas involved with urticaria at baseline and treatment with raniditine or saline
were used as predictors of urticaria presence at 2 these patients, shellfish was implicated
in 14, and fish, fruit, and vegetables were implicated in 9 patients each. Nut or peanut
allergic reactions were suspected in 5 patients. Allergic reactions due to drug
administration was suspected in 28 patients. Of these, anti-infectives were suspected in
11, non-steroidal anti-inflammatory drugs in 8, and psychiatric/neurologic drugs in 4. The
causative agent in 5 patients was uncertain as to whether food or medication related to the
allergic syndrome.
Additional treatments were given to the study patients as detailed in Table II. There was
no difference in the proportion of patients who received intravenous fluids between the 2
groups (data not shown). There were significantly more patients in the placebo group
who received extra antihistamines. Seven patients were admitted to the hospital (3ranitidine, 4-placebo). No adverse effects from the protocol treatment were observed.
The proportions of patients with urticaria at baseline were similar between the 2
groups. For those patients with baseline urticaria, the numbers of areas involved with
urticaria was not significantly different between the 2 groups. The number of patients
with urticaria decreased somewhat more in the ranitidine group at 1 hour. The proportion
of patients without urticaria was significantly greater in the ranitidine group (91.7%) at 2
hours compared to the placebo group (73.8%) (p=0.2). Furthermore the number of areas
10
involved with urticaria at 2 hours was significantly less in the raniditine group compared
to the placebo group (Wilcoxon sign rank test, pO.02). In multivariate modeling, the
number of areas involved with urticaria at baseline and treatment with raniditine or saline
were used as predictors of urticaria presence at 2 proportion of baseline angioedema
cases in the placebo group (Table III). Laryngoedema present at baseline resolved by 2
hours in 4/4 raniditine patients and 2/5 placebo patients.
There was a significant decrease in heart rates over time in both groups (Table
III). A greater decrease in heart rates in the ranitidine group was noted from baseline to
the first hour and 1 hour heart rates were lower in the active group compared to the
placebo group. Analysis of covariance was used further examine heart rate changes
associated with treatment. A significant correlation was observed between the extent of
either baseline erythema and baseline heart rates (Spearman rho test, p<O.OO1). When
baseline erythema extent and epinephrine use were used as covariates (both showing
highly significant covariate effects, p<=0.007 there was a significant treatment by time
effect on heart (p=0.05, baseline contrasted with subsequent values). Post-hoc analysis
contrasting the 1 hour heart rates with or without adjustment for covariates showed a
significant (p<0.05) difference between the 2 groups with lower values for the ranitidine
group, consistent with an relatively more rapid initial decline in heart rates for the
raniditine group.
Patients reported a various symptoms within the first hour of their reaction.
Itching (n=77) and swelling (n=61) predominated. Feeling flushed and throat tightness
11
were also commonly reported (49 and 40 patients respectively). Nausea, dizziness, head
pressure, and palpitations were reported in 30,27, 22, and 20 patients respectively. A
metallic taste, headache, tongue swelling and blurry vision were reported by 15,14,11 and
9 patients respectively. Symptom scores did not show differences over time between the
2 groups.
12
Discussion:
This study demonstrates the additional benefit of adding H2 blockers to H1 antihistamine
treatment in on-going allergic reactions. This finding is important since all but one
previous study have shown only a pretreatment effect for H2 blockade in allergic
reactions. Kaliner et al (6) had demonstrated blunting some of infused histamine’s effects
on normal volunteers using combined H1 and H2 blockade. This isolated mediator
model, probably does not truly simulate in-vivo immediate hypersensitivity reactions,
where multiple mediators are probably released. Differential antihistaminic drug effects
on skin test wheal production by histamine verus allergen attest to the potential dangers
in extrapolating about drug effects with isolated mediator challenges (11). Also in the
model of infused histamine, urticaria and angioedema did not occur (6). These are
cutaneous hallmarks of immediate hypersensitivity reactions, which in our present study
showed effects from combined H1 and H2 treatment.
One question in our study, which was not answered, was whether or not H2 blockade is
beneficial in anaphylaxis. Only 2 patients had hypotension. Twelve patients without
asthma had wheezing. Clearly this subset of patients is too small to examine potential
treatment effects. When patients with more severe manifestations such as upper airway
embarrassment, wheezing in the absence of a history of asthma, and hypotension were
grouped, there was no group effect in the models predicting the primary outcomes (data
not shown). There did not appear to be any differences between patients in this group and
other patients with regards to resolution of respiratory signs over time. Wheezing
resolved in all but one patient by 2 hours and thus a treatment effect on bronchospasm
resolution is difficult to ascertain. Thus although it appears that H2 blockers added to H1
13
antihistamines improve cutaneous manifestations and slowing of heart rates in acute
allergic syndromes, their added benefit in reversing hypotension and bronchospasm
associated with anaphylaxis remains unproven. We were quite interested in the difference
in slowing of heart rates between the 2 groups as this is consistent with a possible effect
on the cardiovascular system from ranitidine (12). It is conceivable that the resolution of
cutaneous edema was accompanied by less vascular permeability and subsequent heart
rate declines in the ranitidine. To our knowledge this study is the first prospective study
to show differential changes in cardiovascular parameters associated with antihistamine
treatment (combined versus H1 alone) in ongoing allergic reactions. As heart rates were
not the primary outcomes, some of the 2 hour values were not recorded. This may have
reduced the power of our study to find a treatment effect. Heart rates may be affected by
factors other than histamine mediated vasodilatation and reflex tachycardia (13). Clearly,
larger studies on more patients with these more serious manifestations need to be
performed in order to test the hypothesis that there is a clinical effect of combined H1 and
H2 blockade in anaphylaxis. In view of the safety and apparently rapid onset of effects,
there would appear to be little reason to not use H2 blockers in patients with anaphylaxis,
since the cutaneous manifestations in these patients, especially urticaria and associated
angioedema, as well as heart rates respond.
Our findings in acute allergic syndromes differ somewhat from the study of Runge et al
(7), who found a post-hoc advantage of addng cimetidine to diphenhydramine for
“clinical relief” of urticaria, as well as improved symptom scores alter 30 minutes of
follow-up but no changes in heart rates. Their study examined only 12 patients who had
combined H1 and H2 blockade as I of 3 treatments (H1 blockade, H2 blockade, or both).
14
Their methodology was suboptimal in 3 important areas. First, there was no statistical
adjustment for multiple comparisons (which given 3 treatment groups and 3 pairwise
comparisons, would have increased the p-values obtained). Secondly baseline
manifestations (which differed significantly between the treatment groups) were not
taken into account in analyzing outcomes. Thirdly an arbitrary definition of “clinical
relief” was used as the main endpoint. Unlike our study, the clear endpoint of resolution
of urticaria with or without angioedema was not studied. Our larger study using ranitidine
as the H2 blocker in 48 patients found a significant difference in resolution of urticaria
(an unambiguous outcome) which remained alter adjustment of baseline urticaria extent,
but not in symptoms. This may in part have been due to a longer time-frame of
observation. It is conceivable that earlier symptom benefits from combination therapy
were eliminated with longer follow-up due to more cumulative symptom resolution in
both groups. The fact that there was more resolution of urticaria in the ranitidine group
has potential implications in terms of resource utilization in emergency departments.
With the quicker resolution of urticaria with or without angioedema, there is a possibility
that patients could be discharged sooner from the ED.
The patients in our study had additional treatments including steroids,
epinephrine, albuterol, and extra doses on antihistamines. The placebo group had
significantly more additional diphenhydramine treatments. This could have potentially
confounded the effect on urticaria resolution if this had been occurred instead in the
ranitidine group. We feel it is likely that more diphenhydramine was given in the placebo
group because the study physicians felt these patients were doing as well. Despite this
additional antihistamine treatment, the placebo group had less resolution of urticaria with
15
or without angioedema. Unlike the situation with asthma (14), we did not find a
significant effect of parenteral corticosteroids in the resolution of allergic cutaneous
manifestations. Unexpectedly, we also did not observe a covariate effect for epinephrine
administration with respect to urticaria resolution. This may have been due to that fact
that clinicians gave both steroids and epinephrine in sicker patients, who are less likely to
respond.
The patients in this study did not have a history of chronic urticaria. Two patients
of several patients who were seen by one of the study physicians after the protocol
treatment, did report recurrences of urticaria. Thus it is possible that a few of the patients
presented with their initial episode of a chronic process. In chronic urticaria, the response
to combined H1 and H2 blockade has been modest at best (15-18). To our knowledge no
large studies have been performed to examine for a benefit in this treatment approach.
Moreover the endpoints for treatment responses in chronic urticaria often relate not to
individual outbreaks of urticaria but to frequency of urticaria and degree of itch over a
longer period of time such as 1-4 weeks. It is likely that these chronic urticaria patients
have a different pathway of cutaneous mast cell release that perhaps may not involve
systemic histamine release as is seen often in more severe acute allergic syndromes. It
thus may not be advisable for clinicians to extrapolate from the treatment effects
observed in this study to assuming potential benefits in patients with chronic urticaria.
16
References:
1.
Schoning 5, Lorenz W. Prevention of allergoid (cutaneous anaphylactoid)
reactions to polygeline (Haemaccel) in orthopaedic patients by premedication with H1and H2-receptor antagonists. Dev Biol Stand 1980;48:241-9.
2.
Lorenz W, Doenicke A, Schoning B, Mamorski J, Weber D, Hinterlang S.
Schwarz B, Neugebauer E. H1 + H2-receptor antagonists for premedication in
anaesthesia and surgery: a critical view based on randomized clinical trials with
Haemaccel and various antiallergic drugs. Agents Actions 1980;1O( Pt
2):114-24
3.
Tryba M, Zevounou F, Zenz M. [Prevention of anaphylactoid reactions using
intramuscular promethazine and cimetidine. Studies of a histamine infusion model]
[Article in German] Anaesthesist 1984;33(5):218-23
4.
Ring J, Rothenberger KH, Clauss W. Prevention of anaphylactoid reactions after
radiographic contrast media infusion by combined histamine H1- and H2-receptor
antagonists: results of a prospective controlled trial. Int Arch Allergy Appi Immunol I
985;78(i):9-14.
5.
Lieberman P. The use of antihistamines in the prevention and treatment of
anaphylaxis and anaphylactoid reactions. J Allergy Clin Immunol 1990;86(4 Pt 2):684-6
6.
Kaliner M, Sigler R, Summers R, Shelhamer JR. Effects of infused histamine:
analysis of the effects of H-I and H-2 histamine receptor antagonists on cardiovascular
and pulmonary responses. J Allergy Clin Immunol 1981 ;68(5):365-71
7.
Runge JW, Martinez JC, Caravati EM, Williamson SC, Hartsell SC. Histamine
antagonists in the treatment of acute allergic reactions. Ann Emerg Med 1992;21 (3)
:237-42
8.
Tabachnick BC, Fidell IS. Using multivariate statistics. HarperCollins, New York,
1989, pp 48-122.
9.
Cody EP and Smith JK. Applied statistics and the SAS programming language.
North-Holland, New York, 1987, pp 132-177.
10.
Ludbrook J. Repeated measurements and multiple comparisons in cardiovascular
research. Cardiovascular Research 1994; 28:303-311.
11.
Monroe FW, Daly AF, Shaihoub RF. Appraisal of the validity of histamineinduced wheal and flare to predict the clinical efficacy of antihistamines. J Allergy Clin
Immunol 1997;99(2) :578-806.
17
12.
Dachman WD, Bedarida C, Blaschke TF, Hoffman BB. Histamine-induced
venodilation in human beings involves both H1 and H2 receptor subtypes. J Allergy Clin
Immunol 1994;93(3):606-14
13.
White MV, Kaliner MA. Histamine in allergic disease in Histamine and H1receptor antagonists in allergic disease(Clinical Allergy and Immunology; 7), editor
Simons FEE. Marcel Dekker, New York 1996; pp. 61-90.
14.
Lin RY, Pesola GP, Bakaichuk L, Heyl CT, Dow AM, Tennenbaum C, Curry A,
Westfal RE. Rapid improvement of peak flow in emergency department asthmatics
treated with parenteral methylprednisolone: a randomized controlled study. Ann
Emergency Med 1999;33:487-495.
15.
OilIer C, Orfanos CE. Z [Management of idiopathic urticaria with H1 + H2
antagonists. A crossover double blind long-term study [Article in German] Hautkr
1983;58(11):785-93
16.
Gay C, Drouet M, Bonneau JC, Le Sellin J, Sabbah A [Bibliographic review of
antihistaminics H2 in the treatment of chronic (or recurrent) idiopathic urticaria].[Article
in French] . Allerg Immunol (Paris) 1988;20(10):377-8
17.
Sharpe CR, Shuster S. In dermographic urticaria H2 receptor antagonists have a
small but therapeutically irrelevant additional effect compared with H1 antagonists alone.
Br J Dermatol 1993;129(5):575-9.
18.
Bodeker PE. Treatment of chronic urticaria with terfenadine and raniditine. A
randomized double-bind study in 45 patients. Eur J Clin Pharmacol. 1986;31 (3):277-80.
18
19
20
21