Implications for practice

Collectively, RCTs done to date fail to show convincing evidence of a clear benefit for oral antihistamines, regardless of whether sedative or non-sedating treatments are used. An ongoing systematic review of these trials conducted by individuals within the Cochrane Skin Group may reveal more precise conclusions if data from similar studies can be combined. In relation to the child described in the case scenario, we would not recommend the use of oral antihistamines except for very occasional use as a sedative (in which case other sedatives might be just as good) and the aggregated systematic

Do oral antihistamines help?

Authors Main reported results Authors' conclusions Comment and quality

Berth-Jones etat. 19891

No evidence of any difference between terfenadlne and

There was no benefit from

Method and concealment of randomisation unclear, study

placebo. Mean scores for Itch over the last 4 days of


described as double blind. Four withdrawals for failure to

treatment (± SEM): 23-95 (± 4-9) for the terfenadlne phase

comply, but not clear at which point and In which Initial arm.

and 25.13 (± 5-1) for the placebo phase. No evidence of

No ITT analysis. Low power to detect carry-over effect In

carry over or period effect when pruritus scores were

only 20 patients. No data In graphical form. First and last


period data not presented separately. Terfenadine double

normal dose

Doherty etat. 19892

Acrlvastine significantly reduced Itching compared with

Acrlvastine and terfenadlne can

Method and concealment of randomisation unclear, study

placebo according to the doctor's assessment (P = 0-021).

partially relieve Itching In atopic

described as double blind. VAS data only given for day 7.

Both acrlvastine (P = 0-026) and terfenadlne (P= 0-037)


Five dropouts (4 active, 1 placebo), no ITT analysis. Unclear

Improved the patient's condition significantly more than

what was being assessed and what was meant by "careful

placebo according to the patient's assessment of the

examination of the skin"

degree of benefit obtained. No significant differences were

found between the two active treatments

Foulds and MacKie

Although there was a significant difference between

Adding an H2 antagonist to the

Method and concealment of randomisation unclear, study


Individual patients for patient-assessed day pruritus

H1 antagonist commonly used

described as double blind. Only one loss to follow up, no ITT

(P<0-001) and night pruritus (0-01<P<0-025), there was no

in young adults with severe

analysis. No actual data given for clinical outcomes; only P

difference between the treatment periods

chronic atopic eczema Is of no

values for statistical comparisons

significant advantage.

Frosch etat. 19844*+

No significant difference between clmetidlne plus

The combination of H1 and H2

Randomisation conducted according to a Latin square,

chlorpheniramine and placebo plus chlorpheniramine for

antagonists Is of no benefit In

study described as double blind. Baseline Itch not given so

day- and night-time patient-assessed Itch at weeks 2, 3

the treatment of atopic eczema.

change cannot be calculated. No standard errors given.

and 4

Missing baseline data. Two dropouts. No ITT analysis

Hamada etat. 1996*5

Itching score and scratch marks Improved significantly.

The combination of terfenadlne

Different topical steroids used In each intervention; no oral

Physician's "Improved" and "markedly improved" 89.3% In

Ingestion and topical alclometa-

placebo. No ITT analysis

antihlstamine-plus-steroid group compared with 50% In

sone application Is more

topical-steroid-only group

effective than betamethasone



Main reported results

Authors' conclusions

Comment and quality

Hannuksela et ai. 1993^

Non-slgnlflcant difference between groups In Intensity of

The sedation observed was

Method and concealment of randomisation unclear.

patient-assessed pruritus at baseline. All groups Improved

probably partly responsible for

Dropout rate high: 51-20 for side-effects (mainly sedation)

significantly (P= 0-005). Improvement significantly more

relief of pruritus. The authors

and 19 non-compilers (drug group not specified). No ITT

pronounced for cetlrlzlne 40 mg than placebo

suggest that cetlrlzlne has

analysis. Possible benefit of cetlrlzlne when used at four

other properties responsible for

times normal dose, but at the expense of sedation

skin lesion healing

Henz et ai. 19987t

Mean overall % response rates based on physician's

Data underline the low efficacy

Neither drug reduced Itching significantly more than

global score were 36-4%, 25-0% and 27-3% In the

of antihistamines In atopic

placebo. Statistics not given for atopic eczema patients. No

azelastlne, cetlrlzlne and placebo groups, respectively.


description of what constituted a response. Placebo looks

Baseline data and exact numbers of atopic eczema

very Impressive, clearly no difference In atopic eczema

patients In each group were not stated. Mean Itching

patients. High dropout rate of 37. No ITT analysis

score decreased from 2-2 to 1-4 In the cetlrlzlne group

and from 2-2 to 1-2 In both azelastlne and placebo groups

(estimated from graphs)

Hjorth 19883

Terfenadlne reduced severity of Itch In approximately 52%

Terfenadlne Is of value In some

No outcome data given and no Information on placebo

of patients, 34% reported no change and 14% reported

patients with atopic eczema

response. Method and concealment of randomisation

Increased severity of Itch. No data given for placebo

and a history of contact

unclear, study described as double blind. Unclear If any


dropouts or withdrawals. Author now deceased

Ishlbashl etat. 19893

No significant difference In general Improvement rating,

No translated data available

No translated data available

(Japanese translation)

overall severity rating and general usefulness rating

among the three dose groups of azelastlne. A significant

difference In Improvement ratio was found In three dose

groups for the signs of Itch, papules, erythema and


Ishlbashl etat. 198910

No difference In final general Improvement rating or

No translated data available

No translated data available

(Japanese translation)

general usefulness rating between the three dose groups

of azelastlne. Effectiveness and usefulness In the

treatment of atopic eczema was considered similar for the

Authors Main reported results Authors' conclusions Comment and quality

three groups. Overall safety rating was significantly higher

In the 4 mg/day group than in the 2 mg/day group. Overall

safety rating similar between the 4 mg/day and ketotlfen

groups and the 2 mg/day and ketotifen groups

Klein and Galant 198011t

Hydroxyzine group had a daytime percentage

Hydroxyzine may be more

Unstable data shown on a graph with Inflationary

Improvement of 32-14 ± 4-98 (mean ± SEM) over baseline

effective than cyproheptadine

percentage scale but no actual data given. Method and

pruritis for the entire week, which Is significantly greater

for the management of pruritus

concealment of randomisation unclear, study described as

(P<0-001) than the percentage Improvement for the

associated with atopic eczema

double blind. Not clear if any dropouts or withdrawals

cyproheptadine group (6-21 ± 4-90)

In children

Langeland et al. 199412

Significant effect of loratadine versus placebo on patient-

Loratadine may be tried as an

Complex design - six consecutive crossovers. Changes In

assessed pruritis during the day and night, and on

adjuvant therapy In the manage

pruritus on VAS all small. No data for period or carryover

severity of rash

ment of severe and moderate

effects shown. Method and concealment of randomisation

atopic eczema, In patients

unclear, study described as double blind (block

complaining of pruritus


LaRosa 1994"

Patient diary card scores showed a statistically significant

The results of this preliminary

Method and concealment of randomisation unclear, study

decrease In erythema and other cutaneous symptoms

study suggest that cetirizine

described as double blind. Only one dropout (voluntary

such as llcheniflcation In the cetirlzine group. Baseline

can effectively control pruritus

withdrawal). Higher baseline scores In those on active

total mean global score reduced from 230 at baseline to

and other cutaneous

treatments suggests that regression to means could partly

155 after 8 weeks' treatment with cetirizine, and from 205

symptoms In children suffering

amount for results

at baseline to 180 after 8 weeks' treatment with placebo

from atopic eczema, without

(estimated from graph)

noticeable side-effects

Monroe 1992"

Dally pruritus score decreased by 57% in the 14 patients

Loratadine demonstrates a

Patients excluded if unresponsive to antihistamines. No

treated with loratadine, 38% In the 14 patients treated with

significant antipruritic effect In

baseline values given. Method and concealment of

hydroxyzine, and 33% In the 13 placebo patients

atopic eczema.

randomisation unclear, study described as double blind.

Very short study (1 week)

Patel etat. 1997m

Loratadine reduced patient-perceived severity of overall

In the management of

Study excluded non-responders before study started:

condition by 20-8% at endpolnt. Incidence of somnolence

symptoms of atopic eczema,

number not reported. Report suggests ITT but failed to

9% with cetirizine and 3% with loratadine

loratadine Is as effective as

Authors Main reported results Authors' conclusions Comment and quality

cetirlzine and Is less sedating

carry It out. Unclear If either drug Is of benefit In absence

of placebo group. Method and concealment of

randomisation unclear, study described as double blind

Savin et al. 197916

Neither drug altered likelihood of scratching bouts

Both trimeprazlne and

Unclear if parallel or crossover study. Length of study

beginning in wakefulness or in any stage of sleep.

trimlpramine resulted in less

unclear. Method and concealment of randomisation

However, sleep was less broken with both drugs, but

broken sleep, with a lessening

unclear, study described as double blind. Withdrawals or

particularly trlmeprazine, and the reduced time spent in

of time awake and in stage-1

dropouts not mentioned. Unclear If changes in sleep

stage-1 sleep accounted for a modest reduction In the

sleep. These actions, which

pattern helped the patient's eczema

overall amount of scratching during the night

may be helpful to some

patients, were associated with

a modest reduction In the

number and length of

scratching bouts

Savin et al. 198617

No significant difference between limb movement

No significant suppression of

No actual data given. Method and concealment of

(measured by limb movement meters) on placebo and

scratching or Itching was

randomisation unclear, study described as double blind.

on active treatment with LN2974. Difference between the


Unclear if any withdrawals or dropouts

mean scores of the VAS assessment of Itching on placebo

and on LN2974 were not significant but favoured LN2974

Simons et al. 198413

Scores for atopic eczema severity and distribution were

Hydroxyzine 0-7 mg/kg three

This trial was buried In the middle of a case series. Only

significantly reduced at the end of treatment for both

times daily was as effective as

presented mean score at end of treatment rather than

doses of hydroxyzine (P<0.05)

hydroxyzine 1-4 mg/kg three

mean change in score. Itch data and baseline scores not

times dally In relieving pruritus

given. Sample size unclear. Point of randomisation was

and promoting resolution of

after the single-dose study. Method and concealment of

skin lesions

randomisation unclear, study described as double blind.

Unclear if any withdrawals or dropouts. Far too small a

study to establish equivalence effects

Simons and Estelle 1999"

During the 18-month long study, only 48 children treated

The safety of cetlrizlne has

Well reported study. Good description of randomisation

with cetlrizlne and 51 treated with placebo dropped out

been confirmed In this

and blinding. Ninety-nine dropouts. No ITT analysis


Main reported results

Authors' conclusions

Comment and quality

Wahlgren et ai. 19902C

Zuluaga de Cadena et al. 198921 (Colombia, translated)

"for any reason" (P= 0-737), only 11 and 15 children, respectively, dropped because of symptoms or events (P= 0-421). Serious events were reported In 37 children (9-3%) receiving cetlrizlne and In 54 children (13-6%) receiving placebo. Hospitalisations were reported In 36 children receiving cetlrizlne and In 47 receiving placebo (P= 0-189). Fatigue and Insomnia were slightly raised In the cetirlzine group. There was no difference In somnolence between cetlrizlne and placebo. No efficacy data were given

No significant changes In difference In itch Intensity between the three treatment periods were detected with Pain-Track, and there was no difference In time awake without pruritus. No significant changes In itch magnitude appeared during each period (days 0-3)

At the end of the 4-week evaluation period, 8 out of 15 patients on terfenadlne compared with 8 out of 17 patients on astemlzole and 6 out of 8 patients on hydroxyzine had Improvement In itch. Global Improvement was noticed In 14 out of 15, 15 out of 17 and 7 out 8 cases, respectively. Improvements In other outcome measures were similar In groups. Other laboratory measures were also recorded prospective study, the largest and longest randomised, double blind, placebo-controlled safety investigation of any H1 antagonist ever conducted In children and the longest prospective safety study of any H1-antagonlst conducted In any age group

The antipruritus effect of 3 days' treatment with terfenadlne (non-sedative) and clemastine (sedative) did not differ from that found with placebo

Antihistamines may be beneficial In atopic eczema, with astemlzole conferring a prolonged benefit

Efficacy outcome data eagerly awaited.

Very short trial (3 days). Method and concealment of randomisation unclear, study described as double blind. No withdrawals or dropouts. Underpowered

Outcome measures and their combinations quite complex. Method and concealment of randomisation unclear. Study described as single (Investigative) blind. No ITT analysis. Small numbers and no placebo group

"used physician-assessed global severity

+used patient-assessed global severity

ITT, intention to treat; VAS, visual analogue scale review is unlikely to be specific enough to inform 7.

this particular process.

0 0

Post a comment