Jump to: Page Content, Site Navigation, Site Search,
You are seeing this message because your web browser does not support basic web standards. Find out more about why this message is appearing and what you can do to make your experience on this site better.
a Heymans Institute of Pharmacology, University of Ghent, De Pintelaan, 185, B-9000 Ghent, Belgium
Correspondence to: Dr Vander Stichele.
Abstract
Objectives: To collect and evaluate all trials on clinical efficacy of topical treatments for head lice.
Design: Systematic review of randomised trials identified from following data sources: Medline, International Pharmaceutical Abstracts, Science Citation Index, letters to key authors and companies, and hand search of journals.
Setting: Trials in schools or communities.
Subjects: Patients infested with lice.
Main outcome measure: Cure rate (absence of live lice and viable nits) on day 14 after treatment.
Results: Total of 28 trials were identified and evaluated according to eight general and 18 lice specific criteria. Of the 14 trials rated as having low to moderate risk of bias, seven were selected as they used the main outcome measure. These seven trials described 21 evaluations of eight different compounds and placebo (all but two evaluations were of single applications). Only permethrin 1% creme rinse showed efficacy in more than two studies with the lower 95% confidence limit of cure rate above 90%.
Conclusions: Only for permethrin has sufficient evidence been published to show efficacy. Less expensive treatments such as malathion and carbaryl need more evidence of efficacy. Lindane and the natural pyrethrines are not sufficiently effective to justify their use.
|
Key messages
|
Introduction
Head lice are among the most common of human ectoparasites, though they are not vectors of serious diseases and in many cases do not cause symptoms.1 2 3 Treatment with natural pyrethrines has been known for more than 100 years, and lindane has been used since the second world war. The synthetic pyrethrines were marketed in the 1950s, malathion and carbaryl in the '60s, and permethrin in the '80s. Although products abound, the prevalence of head lice remains high and epidemics occur regularly despite all efforts at control.3 4 5 Problems such as fear of insects (entomophobia), fear of stigmatisation, and denial of infection by patients and schools may cause under-treatment, overtreatment, and unnecessary prophylaxis, which can lead to development of resistance and insufficient control of epidemics.6 Furthermore, many of the commercially available treatments might be underdosed, incorrectly labelled, or ineffective.
Our aim was to collect and evaluate all trials of clinical efficacy of topical treatments for head lice.7 8
Methods
We searched for trials of topical treatments for people infested with head lice (Pediculus humanus capitis) in which the outcome was measured clinically by inspection of the scalp to determine cure rate (absence of live lice and viable nits).
SEARCH STRATEGY
We searched the medical literature in Medline (1966 to March 1995 using the MESH keywords "Pediculosis," "Lice," "Pediculus"), in International Pharmaceutical Abstracts, and in the Science Citation Index without restriction for language of publication. We scanned the references of all identified clinical trials. We sent letters requesting information about unpublished studies to seven key authors in the subject, to the pharmaceutical companies active in this subject, and to the World Health Organisation centre Vector Biology Control. We hand searched journals in which key references were published for comments, letters, or corrections in the year after publication of the key reference.
REVIEW
We focused on clinical efficacy--the result of pediculicidity, ovicidity, and residual activity--and so we chose cure rate as the main outcome measure for clinical evaluation. The cure rate is the percentage of patients cured after application of the treatment (the 95% confidence interval=p +/- 1.96 (square root p (100-p)/n), where p is the sample percentage and n the number of subjects in the study). Determination of the cure rate by experienced evaluators--on the basis of visual inspection for viable nymphs in nits, hatching nymphs, and adult lice (with a x 10 magnifying lens)--has an acceptable specificity and sensitivity.9 We considered an interval of 14 days between treatment and evaluation to be optimal as this would allow evaluation of the combined effect of pediculicidity (on living lice), ovicidity (on ripening eggs), and residual activity (on hatching nymphs and reinfesting lice).
We evaluated all identified clinical trials with regard to eight general criteria of quality in clinical trials (adapted from Chalmers et al10) and 18 criteria specific for head lice treatment (see table I). We developed these specific criteria after studying the literature to systematically screen the trials for flaws in design, execution, or reporting. Firstly, we made a structured abstract of each clinical trial according to recommended guidelines.11 Each of us then independently assessed the trials. Trials were rejected if four or more flaws in general criteria or 12 or more flaws in treatment specific criteria were found. The remaining trials were rated as having a low risk of bias if less than eight specific criteria were flagged, or a moderate risk of bias if otherwise. Again, we set these cut off points for rating of quality after studying the literature. The structured abstracts, assessment scores, and overall ratings were submitted to an advisory panel (four physicians and a community pharmacist) and discussed until consensus was reached. Of the trials of acceptable methodological quality, we selected those in which the main outcome measure was the cure rate at 14 days after treatment.
TABLE I--Criteria of quality in evaluation of design, execution, and
reporting of trials on clinical efficacy of topical treatment for head lice
------------------------------------------------------------------------------
Item No Criterion
------------------------------------------------------------------------------
General
1 Randomisation procedure
2 Concealment of allocation to patients
3 Concealment of allocation to investigators
4 Precision of definition of exclusion criteria
5 Handling and reporting of drop outs
6 Ethical procedures*
7 Statistical procedures
8 Appropriateness of conclusions
Treatment specific
1 Documentation of prior exposure of screened population to
pesticides (therapeutic or agricultural)
2 Documentation of history of previous lice treatment and
comorbidity of index patients
3 Quality of informed consent procedure (involvement of parents)
4 Inclusion criteria (definition of "current" head lice infection)
5 Specification of the formulation of active ingredients
6 Storage and manipulation of pharmaceutical compound
7 Time and season of study
8 Prevalence of lice in study area
9 Standardisation of cotreatment and swimming
10 Identity of applicants and evaluators
11 Application procedure
12 Intensity of tracing contacts
13 Use of nit combs
14 Documentation of pediculicidity
15 Documentation of ovicidity
16 Documentation of residual activity
17 Documentation of cure rate
18 Adverse events reported
------------------------------------------------------------------------------
*More relevant to good clinical practice than evaluation of bias. |
Results
SELECTION OF TRIALS
We identified 28 trials of clinical efficacy, 27 from the computer databases12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 and one supplied by an author in reply to our request.39 We also identified an internal document of the Wellcome company describing a series of 11 small (11<n<74) unpublished comparative trials of permethrin and malathion performed between 1983 and 1986 and apparently yielding high cure rates for permethrin and malathion.40 As we decided not to accept the company's demands for confidentiality, the full texts of these studies were not made available to us and these trials were not included in the analysis. Hand searching did not reveal additional relevant reports of trials. The search in the Science Citation Index showed only limited citing activity in this field. An official of WHO confirmed that the debate on the choice of head lice treatment for the list of essential drugs had been based on expert opinion without a formal literature review. Narrative reports of studies and older reports of treatment campaigns41 42 43 44 45 were not included in the analysis.
Four studies were not controlled,12 18 26 39 three studies (two of malathion and one of permethrin) were placebo controlled,15 21 23 and the remaining 21 studies were comparisons between two or more active substances. We rated the quality of the identified trials according to the criteria in table I and rejected 14 of the trials because of an excess of general or treatment specific flaws (table II). Seven trials were excluded from the analysis, although their methodological quality was acceptable, because the cure rate was not determined on day 14 after treatment (three measured cure rate on day 7,15 21 38 and four measured it on day 21 or later16 29 30 32). The characteristics of the excluded trials are listed in table III.
TABLE II--Assessment of 28 trials according to eight general and 18 treatment specific criteria of quality
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------
General item No* Treatment specific item No*
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Study 1 2 3 4 5 6 7 8 Total 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Total Risk of bias
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Trials of acceptable quality with cure rate at day 14
Maunder14 F F F 3 F F f F F F F f f 9 Moderate
Brandenburg et al22 F F F 3 F F F F f 5 Low
Taplin et al23 F 1 f f f 3 Low
Bowerman et al24 F F 2 F f F F f f f 7 Low
Carson et al27 F F 2 f F f f f f f 7 Low
DiNapoli et al28 F F F 3 F f f F f f f 7 Low
Clore et al36 F F 2 F f F F f f f f f 9 Moderate
Trials of acceptable quality without cure rate at day 14
Taplin et al15 F 1 F F F F f f f 7 Low
Mathias et al16 F F F 3 F F F F f f f f 8 Moderate
Urcuyo et al21 F F 2 F F F F f f f f 8 Moderate
Miller et al29 F 1 F F F F f F f F f f 10 Moderate
Kyle30 F F 2 f f F F f f F f f f 10 Moderate
Sexton et al32 F F 2 f f F F F f f F f f f 11 Moderate
Chosidow et al38 F F 2 f F F f f F f 7 Low
Trials of unacceptable quality
Blommers et al12 F F F F 4 F F F F F F F F F F f F F F 14 High
Preston et al13 F F F F 4 F F f F F f f f F F F f 12 High
De Boer17 F F F F F 5 F f F F F F F f f f f f f 13 High
De Boer et al18 F F F F F 5 F F f F F f f f F f f f 12 High
Mazas et al19 F F F F 4 f F F F F F f F f F f f F 13 High
Donaldson et al20 F F 2 F f F f F f f F f f f f f 13 High
Armoni et al25 F F F F 4 F F f F F F f f f f 10 High
Mazas et al26 F F F F 4 F F F f F F f F f F f f f 13 High
Rousset et al39 F F F F F 5 F F F f F F F F F f f 11 High
Mathias et al31 F F F F 4 F F f F f F F f 8 High
Doss et al33 F F 2 f f F F F F f f F f f f 12 High
Jolley et al34 F 1 f f F F f f F f F f f f 12 High
Fan et al35 F F F F 4 F F F F F F f F f f f f 12 High
Burgess et al37 F F F F F 5 F F f F F f F F 8 High
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------
*See table 1 for details of items.
F=major flaw; f=minor flaw. |
TABLE III-Overview of clinical trials excluded from analysis
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Percentage difference in
No of cure rate from highest
Treatment (single application index Cure outcome (95% confidence
Study unless stated otherwise) patients rate (%) interval)
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Trials of acceptable quality without cure rate at day. 14
Taplin et al (cure rate at day 7)15 Malathion 0.5% lotion 65 95
Control vehicle 47 45 50 (35 to 65)*
Matthias et al (cure rate at day 28-40)16 Malathion 0.5% lotion 29 76
Lindane 1% shampoo 33 78 2 (0 to 23)
Urcuyo et al (cure rate at day 7)21 Malathion 0.5% lotion 61 85
Control vehicle 58 7 78 (67 to 89)*
Miller et al (cure rate at day 21)29 (delta)-Phenothrin 0.2% lotion 32 100
(delta)-Phenothrin 0.2% lotion 24 100 0
Kyle (cure rate at day 21)30 (delta)-Phenothrin 0.2% shampoo 39 87
Malathion 0.5% lotion 38 82 5 (0 to 21)
Sexton et al (cure rate at day 21-28)32 (delta)-Phenothrin 0.2% shampoo 27 96
Carbaryl 0.5% lotion 23 87 9 (0 to 25)
Chosidow et al (cure rate at day 7)38 Malathion 0.5% lotion 94 95
(delta)-Phenothrin lotion 95 39 56 (45 to 67)*
Trials of unacceptable quality
Blommers et al12 Lindane 1% lotion+ 110
Uncontrolled
Preston et al13 Carbaryl 0.5% lotion 5
Carbaryl 1% gel shampoo 26
De Boer17 Malathion 0.5% lotion 51
Bioallethrin 1.8%+butoxide 76
De Boer et al18 Malathion 0.5% lotion 51
Uncontrolled
Mazas et al19 Permethrin 1% lotion 10
Permethrin 1% lotion 10
Donaldson et al20 (delta)-Phenothrin 0.2% shampoo 42
Carbaryl 1.5% gel shampoo 34
Armoni et al25 Pyrethrin 0.3%+butoxide shampoo 50
Pyrethroid 0.66%+butoxide spray 50
Malathion 0.4% lotion 50
Carbaryl 0.6% shampoo 50
Carbaryl 0.5% lotion 50
Mazas et al26 Malathion 0.5% lotion 37
Uncontrolled
Rousset et al39 Bioallethrin 0.66%+butoxide spray 100
Uncontrolled
Mathias et al31 Lindane 1% 25
Pyrethrin 0.3%+butoxide 28
Doss et al33 (delta)-Phenothrin 0.2% lotion 50
Malathion 0.5% lotion 23
Carbaryl 0.5% lotion 28
Jolley et al34 (delta)-Phenothrin 0.2% shampoo 25
Carbaryl 0.5% shampoo 25
Fan et al35 Permethrin 1% creme rinse 529
Bioallethrin 0.66% spray 314
Malathion 1% lotion 519
Lindane 1% powder 249
Burgess et al37 Synergised pyrethrin mousse 42
Permethrin 1% creme rinse 10
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
*Significant difference (zero not included in 95% confidence interval of difference).
+Two applications. |
The remaining seven trials were of acceptable methodological quality and had the cure rate on day 14 as main outcome measure.14 22 23 24 27 28 36 Five of the trials had an overall quality rating of low risk of bias, while two had a moderate risk of bias (table II). Three trials were conducted in an area with high background prevalence of head lice (>50% of screened population).14 23 24
RESULTS OF SELECTED TRIALS
In the seven selected trials 21 individual evaluations of topical treatments were performed, comparing placebo and eight compounds (lindane, bioresmethrin, chlorphenamide, (delta)-phenothrin, pyrethrin, malathion, carbaryl, and permethrin), and all but two evaluations27 36 were of single applications (table IV). We juxtapositioned the results obtained in different trials for the same compounds, and the figure shows the cure rates for each compound. The cure rate with placebo was 6%,23 showing the lack of placebo effect and spontaneous remission with this condition. We found six evaluations of lindane; in none of them did the lower confidence limit of the cure rate exceed 90%, and in two trials even the upper confidence limit was below 90%. For the natural pyrethrines (pyrethrin, bioresmethrin, chlorphenamide, and (delta)-phenothrin), all the evaluations resulted in cure rates with lower confidence limits below 90%. Only one evaluation was available for carbaryl 0.5% lotion and malathion 0.5% lotion, both giving cure rates with lower confidence limits above 90%. We found five evaluations of permethrin 1% creme rinse (in a single application of 10 minutes) with cure rates of nearly 100% and lower 95% confidence limits above 90%. Two of these studies were high quality studies in populations with a high background prevalence of head lice (>50% of screened population infested).23 24
TABLE IV--Overview of included trials
----------------------------------------------------------------------------------------------------------------------------------
Percentage difference in
No of Cure rate (%) cure rate at day 14 from
Ttreatment (single applications index ------------------- highest outcome (95%
Study unless stated otherwise) patients Day 7 Day 14 confidence interval)
----------------------------------------------------------------------------------------------------------------------------------
Maunder14 Carbaryl 0.5% lotion 81 100
Carbaryl 0.5% shampoo 64 97 3 (0 to 7)
Bioresmethrin 0.2% lotion 49 92 8 (0 to 7)
Chlorphenamide 0.2% lotion 93 86 14 (7 to 21)*
Lindane 0.5% lotion 97 91 9 (3 to 15)*
Lindane 1% shampoo 57 86 14 (5 to 23)*
Malathion 0.5% lotion 108 98 2 (0 to 5)
Brandenburg et al22 Permethrin 1% creme rinse 257 99 99
Lindane 1% shampoo 251 92 85 14 (9 to 19)*
Taplin et al23 Permethrin 1% creme rinse 29 100 97
Control vehicle 34 9 6 91 (81 to 100)*
Lindane 1% (non-random) 30 67 43
Bowerman et al24 Permethrin 1% creme rinse 195 99 98
Lindane 1% shampoo 99 90 76 22 (13 to 31)*
Carson et al27 Permethrin 1% creme rinse 27 96 100
Pyrethrin 0.3% lotion+ 31 45 94 6 (0 to 14)
Di Napoli et al28 Permethrin 1% creme rinse 107 98 96
Permethrin 0.3% lotion 106 85 62 34 (24 to 44)*
Clore et al36 Lindane 1% shampoo+ 30 80 93
Permethrin 1% creme rinse 32 91 87 6 (0 to 21)
Pyrethrin 0.3% five brands 31 79 86 7 (0 to 18)
----------------------------------------------------------------------------------------------------------------------------------
*Significant difference (zero not included in 95% confidence interval of difference).
+Two applications. |
We also made an intrastudy comparison of the two largest trials.22 24 Both trials were of high quality and compared single applications of permethrin 1% creme rinse with lindane 1% shampoo. The odds ratio of treatment failure for lindane versus permethrin was 15.11 (95% confidence interval 4.60 to 49.62) in one study22 and was 15.28 (5.13 to 45.52) in the second study.24 After performing a Breslow-Day test for homogeneity of odds ratios (P=0.99), we obtained the Mantel-Haenszel summary odds ratio of 15.18 (7.99 to 28.84). Hence, the risk of treatment failure was likely to be at least eight times higher with lindane than with permethrin.
Discussion
Our aim was to collect details of all trials on the clinical efficacy of treatment for head lice and to describe the results of the trials that were not invalidated by too many flaws. To our surprise, we found only 28 published studies.
METHODS OF REVIEW
We cannot exclude the possibility that important research findings were missed by our method of retrieval. We did not engage in a hand search of core journals, as recommended for structured reviews,46 because there did not seem to be a core group of journals, where research publications are concentrated,47 for this subject. Our failure to obtain the full text of unpublished trials comparing permethrin with malathion excluded evidence for the efficacy of malathion.
In this first systematic approach to the treatment of head lice it was not possible to determine the acceptance criteria a priori. The rating system was instead developed after study of the identified trials. We chose cure rate at 14 days as the main outcome measure since it was the most commonly used criterion for efficacy and is, in our opinion, the most appropriate outcome measure. We preferred not to consider the cure rate at seven days, as hatching of nymphs can take longer than this48 and as a week is too short to evaluate the effect of residual activity on reinfestation (which is important in stopping transmission during epidemics).49 We have, however, presented cure rates at day 7 and beyond day 14 (tables III and IV), but these results do not challenge the overall conclusions of our review.
DATA EXTRACTION
We presented the efficacy data of each active ingredient by juxtaposition of treatment groups from the seven selected trials. This procedure is methodologically weak but has the advantage of extracting at least some of the limited knowledge from the clinical studies. The representation of the confidence intervals in the figure should be interpreted with caution, as many of the results fall in the extreme end of the range of cure rates, where confidence intervals based on the sample percentage tend to shrink to zero.
No of
Control vehicle patients
(Taplin et al23) 34
-------------------------------------------------------------------------------------------------------------------
Lindane
1% Shampoo (Clore et al36)* 30
0.5% Shampoo (Maunder14) 97
1% Shampoo (Maunder14) 57
1% Shampoo (Brandenburg et al22) 251
1% Shampoo (Bowerman et al24) 99
1% Shampoo (Taplin et al23) 30
-------------------------------------------------------------------------------------------------------------------
Pyrethrines
Pyrethrin 0.3% lotion (Carson et al27)* 31
Bioresmethrin 0.2% lotion (Maunder14) 49
Chlorphenamide 0.2% lotion (Maunder14) 93
Pyrethrin 0.3% (five brands)(Clore et al36) 31
-------------------------------------------------------------------------------------------------------------------
Carbaryl
0.5% Lotion (Maunder14) 81
0.5% Shampoo (Maunder14) 64
-------------------------------------------------------------------------------------------------------------------
Malathion
0.5% Lotion (Maunder14) 108
-------------------------------------------------------------------------------------------------------------------
Permethrin
1% Creme rinse (Carson et al27) 27
1% Creme rinse (Brandenburg et al22) 257
1% Creme rinse (Bowerman et al24) 195
1% Creme rinse (Taplin et al23) 29
1% Creme rinse (DiNapoli et al28) 107
1% Creme rinse (Clore et al36) 32
0.3% Lotion (DiNapoli et al28) 106
-------------------------------------------------------------------------------------------------------------------
*Double application 0 10 20 30 40 50 60 70 80 90 100
Cure rate at day 14 (%)
Cure rates (95% confidence intervals) at day 14 after treatment for head lice. (Filled rectangles indicate high
background prevalence (>50% of screened population infested). Products were applied once unless
indicated otherwise) |
Within the seven selected trials it was difficult to make sensible intrastudy comparisons. The sample size of two of the trials27 36 was insufficient for testing relevant differences. Two trials were selected despite flaws in randomisation, one because of the comparison with placebo23 and one because of the range of products tested.14 The relevance of testing a product against itself in an underdosed formulation in one study28 can be questioned. Hence, we limited ourselves to the comparison of the two bigger high quality trials, which compared permethrin with lindane.22 24
OTHER ASPECTS OF EVALUATING TREATMENT
We made no attempt to formally weigh criteria such as side effects, toxicity, and cost. In the course of reviewing the literature, we found only one large scale postmarketing surveillance of safety, which provided evidence for the safety of permethrin.50
Theoretically, products with residual activity might facilitate selection of resistant strains of head lice, and proposals have been made for a rotational or mosaic treatment strategy.49 51 There is no convincing evidence for a need for such a strategy, but development of resistance should be monitored if the therapeutic arsenal diminishes to a few products of proved efficacy.
RECOMMENDATIONS FOR FUTURE TRIALS
The number of well conducted trials of clinical efficacy in this subject of medical and economic importance is limited, and recommendations for treatment published in the medical literature are not sufficiently sustained by the results of research. Our list of treatment specific criteria could be a starting point for evaluation of new trials in the subject. Moreover, inspection of the figure leads us to recommend that only products with an expected cure rate of over 90% should be tested and that this should be done in trials with sufficient power to establish cure rates with a lower confidence limit above 90%. We propose the use of equivalence testing (testing for non-null hypothesis) and of the odds ratio of treatment failure with its 95% confidence interval, as the
2 test might not be valid for testing differences in cure rates near 100%. Future research on head lice treatments should preferably test single applications of compounds, as was the case in almost all the acceptable clinical trials in this study. Any treatment for lice might be effective if it is applied repeatedly over a short interval.52 We found many examples of instructions on labels encouraging multiple application, especially with products that lacked well documented efficacy. In some cases these instructions clearly played on people's entomophobia to stimulate consumption, as has been stated by others.6
We are indebted to the members of the advisory panel: C Carton, army pest control physician; J M Kaufman, endocrinologist; G Laekeman, professor of pharmacy; K Seynaeve, school health services physician; and E Van Hecke, dermatologist. We thank D De Bacquer for statistical advice and A Herxheimer for critical comments on the manuscript.
Funding: This study was supported by a grant from the Belgium Ministry of Health.
Conflict of interest: None.