The authors have declared that no competing interests exist.
Current address: ORISE Public Health Research Fellow, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America
WHO’s Global Programme to Eliminate Lymphatic Filariasis (LF) uses mass drug administration (MDA) of anthelmintic medications to interrupt LF transmission in endemic areas. Recently, a single dose combination of ivermectin (IVM), diethylcarbamazine (DEC), and albendazole (ALB) was shown to be markedly more effective than the standard two-drug regimens (DEC or IVM, plus ALB) for achieving long-term clearance of microfilaremia.
To provide context for the results of a large-scale, international safety trial of MDA using triple drug therapy, we searched Ovid Medline for studies published from 1985–2017 that reported adverse events (AEs) following treatment of LF with IVM, DEC, ALB, or any combination of these medications. Studies that reported AE rates by treatment group were included.
We reviewed 162 published manuscripts, 55 of which met inclusion criteria. Among these, 34 were clinic or hospital-based clinical trials, and 21 were community-based studies. Reported AE rates varied widely. The median AE rate following DEC or IVM treatment was greater than 60% among microfilaremic participants and less than 10% in persons without microfilaremia. The most common AEs reported were fever, headache, myalgia or arthralgia, fatigue, and malaise.
Mild to moderate systemic AEs related to death of microfilariae are common following LF treatment. Post-treatment AEs are transient and rarely severe or serious. Comparison of AE rates from different community studies is difficult due to inconsistent AE reporting, varied infection rates, and varied intensity of follow-up. A more uniform approach for assessing and reporting AEs in LF community treatment studies would be helpful.
WHO’s Global Programme to Eliminate Lymphatic Filariais (LF) supports annual mass drug administration to over 400 million people in LF-endemic areas each year. Two drug combinations (either DEC or ivermectin, given with albendazole) have been recommended in most endemic areas. With the exception of well-described serious adverse events (AEs) occurring in patients with high level loiasis, severe AEs due to these medications are extremely rare. Mild to moderate AEs, however, are common, particularly in patients with active filarial infection. In this manuscript we synthesize published data on AEs following single-dose treatment of LF with ivermectin, DEC, or albendazole. This provides a background against which to compare the safety of triple drug therapy (ivermectin, DEC, and albendazole) recently endorsed by WHO, and provides a useful context for evaluating safety of new treatments for LF. The compiled data illustrate that transient, mild to moderate AEs following single-dose LF treatment are common in microfilaremic patients and are much less common in amicrofilaremic patients. They also show that passive surveillance for post-treatment AEs underestimates AE incidence and suggest that adherence to common reporting standards would improve the usefulness of AE reporting in filariasis studies.
Infection with the filarial nematode parasites
Medications used for MDA include diethylcarbamazine (DEC), ivermectin (IVM) and albendazole (ALB). The combination of IVM plus ALB is used in areas of Africa where onchocerciasis (river blindness) is co-endemic with LF. Twice yearly ALB alone is recommended for LF-endemic areas of Africa that are co-endemic for loiasis, and DEC plus ALB is used in the rest of the world. Serious (life-threatening) AEs due to MDA are exceedingly rare [
A promising new combination therapy for LF that combines a single dose of IVM, DEC, and ALB (IDA) appears to be highly effective [
The primary outcome of interest for the quantitative synthesis was the proportion of participants experiencing at least one AE (aggregate AE rate). Rates of individual AEs were a secondary outcome. We did not use a pre-specified AE definition, but rather accepted all AEs reported by the authors of the individual studies. In this manuscript we classify AEs as mild if they do not interfere with normal daily routine (work or school), as moderate if they interfere with daily routines (work or school) but not with activities of daily living, and as severe if they interfere with activities of daily living or cause temporary incapacitation. These designations correspond to Common Terminology Criteria for AEs grades 1 (mild), 2 (moderate), and 3 (severe). Serious AEs are those that are life-threatening or result in hospitalization or permanent injury (grade 4) or are fatal (grade 5) [
We reviewed AE data from studies of LF treatment with single-dose regimens that were published between 1985 and 2017. We searched Ovid Medline and Embase for any articles with Medical Subject Headings (MeSH) terms “Elephantiasis, Filarial” and “Drug Therapy” plus any of the following terms: “Adverse Events”, “Poisoning”, or “Toxicity”. We limited our search to English or French language manuscripts dealing with human infections. The most recent search was conducted on 21 Aug, 2017. Two authors (PB and CH) reviewed each publication and gathered additional pertinent publications from articles referenced therein. Publications with sufficient AE data were selected for a quantitative analysis of AEs as described below. We did not pre-specify nor register a review protocol. We did not attempt to contact authors to identify additional studies.
Studies published after 1985 that reported AE rates following single-dose LF treatment with IVM, DEC, or ALB (alone or in combination) were included. Studies dealing with multi-day courses (generally of DEC) were reviewed, but excluded from the quantitative analysis, as were studies that either provided inadequate information on AEs by treatment arm or did not conduct AE surveillance within one week following treatment. Complete inclusion and exclusion criteria are shown in
Inclusion Criteria | Exclusion Criteria |
---|---|
• Treatment with DEC, IVM, or ALB |
• Studies of multi-dose treatments |
*We included studies that used a preliminary clearing dose [
From studies meeting inclusion criteria we extracted data including: study location (country), age range and gender of participants, intensity of surveillance, treatment regimen, Mf prevalence, geometric mean Mf counts, presence of co-infections, overall rate of AEs, and rates for any specific AEs reported. For studies that reported AEs following multiple MDA treatment rounds, we included only the AE rates that occurred after the first treatment. For studies in which one but not all treatment arms met inclusion criteria (for example, when single dose IVM or DEC was compared to 12 days of DEC), we included data only from the arm(s) meeting inclusion criteria. The number of participants reported in our analysis is the number for whom AE surveillance was conducted, which was sometimes lower than the total number treated. For example, one study conducted active post-MDA surveillance within a subset of 483 persons among 8 million people treated [
All extracted data were analyzed using Stata version 12.1 (College Station, TX). Because the data were not normally distributed, we report means and interquartile ranges (IQR) and use boxplots for graphical representation. Since AE reporting was insufficiently uniform among included studies, we did not attempt a formal meta-analysis of AE rates, nor did we attempt statistical analyses. Rather, we sought to present a graphical synthesis of data from these disparate studies to illustrate the range of data and an estimate of central tendency (median and interquartile range). To assess for reporting bias in individual studies, we stratified surveillance for AEs in each study as active (individual participants were contacted and asked about AEs) or passive (individuals with AEs had to seek out the study team to report). The quality of active surveillance was further categorized as “high” (at least daily contact during the first 72 hours), “moderate” (at least one contact within first 72 hours), or “low” (participants contacted after 72 hours). Although we hoped to analyze the effect of each extracted variable on reported AE rates, we found that the quality of data reported for most parameters was insufficient. We therefore limited our analysis to an ad hoc comparison of treatment regimens, Mf status, and intensity of AE surveillance.
Many informative articles that reported AEs following treatment for LF could not be included in our quantitative analysis either because they reported composite AE scores rather than rates, or because they did not report AE rates separately by treatment group. We have attempted to review some of the observations from both included and excluded publications in the following paragraphs.
Transient mild to moderate adverse reactions such as fever, headache, dizziness, malaise, myalgia, fatigue, and gastrointestinal upset are common after treatment of LF and are primarily related to dying Mf [
Localized AEs following LF treatment are much less common than systemic AEs. They include the development of subcutaneous or scrotal nodules, spermatic cord swelling, new onset or increased hydrocele, dilated and inflamed lymphatic vessels (the “string sign”) [
The decline in Mf counts following IVM treatment may be somewhat slower in persons with brugian filariasis [
To summarize published rates of AEs following single-dose treatment of LF, and to explore how these might differ by treatment medication and AE surveillance, we compiled data from articles with adequate AE reporting into a combined analysis. Among 162 full text articles reviewed, 55 contained AE data that met criteria for inclusion (
Study location | N Studies | N participants | References |
---|---|---|---|
Hospital or Clinic Based Studies with High Quality, Active Surveillance | |||
Brazil | 3 | 194 | [ |
China | 1 | 60 | [ |
Egypt | 1 | 71 | [ |
French Polynesia | 7 | 314 | [ |
Haiti | 4 | 340 | [ |
India | 11 | 591 | [ |
Indonesia | 1 | 15 | [ |
Malaysia | 1 | 40 | [ |
Papua New Guinea | 1 | 24 | [ |
Senegal | 1 | 16 | [ |
Sri Lanka | 2 | 77 | [ |
Tanzania | 1 | 25 | [ |
Community Based Studies with Active Surveillance | |||
Brazil | 2 | 818 | [ |
Egypt | 1 | 28 | [ |
French Polynesia | 2 | 4,421 | [ |
Ghana | 2 | 1,299 | [ |
India | 5 | 10,596 | [ |
Papua New Guinea | 1 | 966 | [ |
Samoa | 1 | 458 | [ |
Sri Lanka | 1 | 31 | [ |
Tanzania | 1 | 57 | [ |
Community Based Studies with Passive Surveillance | |||
Ghana | 1 | 15,020 | [ |
Haiti | 2 | 74,968 | [ |
Kenya | 1 | 170 | [ |
Mali | 1 | 42 | [ |
As shown in
Groups where no participants (0%) or all participants (100%) were Mf positive were from hospital or clinic-based clinical trials; those with “varied” Mf status were from community studies. A) Results when each study group was considered to have equal weight. B) Results when results were adjusted to consider the number of participants in each study group. Boxes indicate the interquartile range (IQR) and dashed line the median. Whiskers indicate the upper and lower adjacent values: upper = the greatest value less than (75th percentile + (1.5 x IQR)); lower = the lowest value greater than (25th percentile–(1.5 x IQR)).
Aggregated AE rates by treatment regimen from studies with active surveillance are shown in
Groups where no participants (0%) or all participants (100%) were Mf positive were from hospital or clinic-based clinical trials; those with “varied” Mf status were from community studies. The top panels show equal weighting for each study, regardless of participant number; results in the bottom panels are weighted by participant number. Boxes indicate the interquartile range (IQR) and dashed line the median. Whiskers indicate the upper and lower adjacent values: upper = the greatest value less than (75th percentile + (1.5 x IQR)); lower = the lowest value greater than (25th percentile–(1.5 x IQR). Abbreviations: ALB = Albendazole, DA = DEC + Albendazole, DI = DEC + Ivermectin, IA = Ivermectin + Albendazole, IDA = Ivermectin + DEC + Albendazole, IVM = Ivermectin.
Aggregate AE rates after treatment of
Results varied widely in studies with active surveillance that reported specific AEs after treatment of Mf positive persons. Rates of fever were 100% in some studies, and other systemic symptoms including myalgia/arthralgia, dizziness, giddiness, weakness, fatigue, and malaise were also very common. Rates of systemic events in amicrofilaremic individuals were much lower (
Only studies with active AE reporting are included and each study arm is given equal weight (not weighted by participant number). Boxes indicate the interquartile range (IQR) and dashed line the median. Whiskers indicate the upper and lower adjacent values: upper = the greatest value less than (75th percentile + (1.5 x IQR)); lower = the lowest value greater than (25th percentile–(1.5 x IQR)).
DEC and IVM data include all study arms where the indicated drug was given alone or in combination (with or without ALB). For clarity, the ALB panel shows only study arms in which ALB alone was given. The upper panels assign each study are equal weight; the lower panels are weighted according to number of participants per study arm. Boxes indicate the interquartile range (IQR) and horizontal line the median. Whiskers indicate the upper and lower adjacent values: upper = the greatest value less than (75th percentile + (1.5 x IQR)); lower = the lowest value greater than (25th percentile–(1.5 x IQR)).
In this review we initially sought to quantify the effects of various factors on AE rates that occur following MDA for LF. We quickly realized that the heterogeneity in the way AEs have been reported in the literature would not allow a meaningful quantitative multivariate analysis. We nevertheless felt a compilation of reported AE rates would be beneficial. Despite the limitations of combining data from methodologically disparate studies, we believe the compiled data illustrate the following main points: 1) AEs are very common in microfilaremic patients after single-dose treatment of LF with drugs (IVM and DEC) that rapidly reduce Mf counts. 2) AEs are much less common in amicrofilaremic participants, regardless of treatment regimen. 3) Passive surveillance tends to underestimate the occurrence of AEs, and 4) Heterogeneity in the stringency of AE surveillance and format of AE reporting makes comparisons between studies difficult.
The relationship between AE rates and the prevalence of microfilaremia is illustrated by the striking differences between study arms with 100% microfilaremia and those with no microfilaremia. It would have been interesting to compare Mf prevalence to AE rates among the community studies with varied Mf prevalence; this was not attempted because of the variability in AE reporting for these studies and because uncertainty regarding true Mf rates would have made this comparison unreliable.
It is clear that community proclivities for reporting AEs vary from place to place and study to study. This is perhaps most evident in reported AE rates after placebo treatment. Studies with highly active AE surveillance in Haiti and Tahiti reported high AE rates after placebo treatment [
Nearly all the studies cited in this review reported AE rates in some manner, but we were only able to include 55 in the quantitative synthesis. The primary reason for excluding studies was that they did not present AE data in a way that linked AE rates to treatment regimens. For example, many studies reported AE severity scores rather than rates. Others reported that AE rates did not differ significantly between treatment groups, but did not report the numbers for each group. When AE rates were reported by treatment group, comparisons were often hampered by non-standardized AE reporting procedures. Some authors did not report the timeframe over which AE surveillance was conducted, making it difficult to surmise whether early or late AEs may have been missed. Although most studies included in our analysis described whether AE surveillance was active or passive, many contained insufficient detail to determine how sensitive the study procedures were for detecting AEs. For example, ascertainment rates (the proportion of participants in community-based studies who were actually visited and queried about AEs) were almost never reported.
This review has several strengths and weaknesses. The primary strength is that it compiles data from 30 years of published studies. It also illustrates how variable AE reporting can be, and it provides a context for interpreting AE rates observed in future LF treatment studies. One weakness was our inability to include data from many high quality studies that did not report AEs by treatment arm. In addition, because we restricted our analysis to studies of single-dose therapy, many rich and highly informative studies that used multi-dose treatment regimens were excluded. In general, the pattern of AEs reported in such studies was similar to single dose studies. That is to say, the rate and severity of AEs increased with increasing Mf counts and most AEs occurred during the first 48 hours after the initial treatment dose [
The heterogeneity in AE reporting among the reviewed studies highlights the need for a more structured approach to AE reporting in LF treatment studies. Although this problem is not unique to filariasis [
Clearly specify the method for ascertaining AEs.
Was an attempt made to contact each participant to assess for AEs? When and how often were AEs assessed?
Report the ascertainment rate (the percentage of participants assessed for AEs).
Report numerators and denominators.
How many adverse events occurred among how many participants? Report the actual numbers, even if severity scores are used, to allow readers to calculate rates.
Use standardized grading for reporting AE severity.
Follow CONSORT guidelines for reporting harms.
In conclusion, this review has shown that AEs following single dose treatment of LF are common and should be expected in microfilaremic patients. This information provides a useful context for understanding AEs observed with new treatments for LF. Clear and detailed reporting of AEs in community treatment studies is essential to accurately inform elimination program workers and their communities, and to set appropriate expectations. The fear of MDA-associated AEs is often out of proportion to the actual risk, because most post-treatment AEs are mild and transient. A frank explanation of AEs as a marker for treatment efficacy by program managers and community health workers may improve compliance with MDA and facilitate LF elimination efforts.
(DOC)
(XLSX)
(DOC)
We would like to thank Kim Lipsey, Medical Librarian at the Bernard Becker Medical Library, for her assistance with developing the initial literature search strategy and conducting the literature searches.