A meta-analysis comparing the safety and efficacy of azithromycin over the alternate drugs used for treatment of uncomplicated enteric feverNA Trivedi, PC Shah
Department of Pharmacology, Medical College, Baroda, Gujarat, India
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0022-3859.97172
Source of Support: None, Conflict of Interest: None
Background: Drug-resistant typhoid fever is a major clinical problem globally. Emergence of multidrug-resistant (MDR) S. Typhi has complicated therapy by limiting treatment options. Objectives: A meta-analysis was planned to determine the strength of evidence supporting use of azithromycin over the alternate drugs available for treatment of uncomplicated typhoid fever. Materials and Methods: Studies were identified using electronic database such as MEDLINE and other data at the National Library of Medicine assessed using PUBMED search engine as well as Cochrane Clinical Trial Register. Randomized control trials (RCTs) comparing azithromycin with chloramphenicol, fluoroquinolones and cephalosporins in culture-proven enteric fever were included. Data was extracted and methodological quality was assessed. Risk ratio (RR) with 95% confidence intervals was estimated for the dichotomous outcomes and mean difference (MD) with 95% confidence was estimated for continuous data. Primary outcomes studied were clinical failure (CF), microbiological failure, and relapse. Results: A total of seven RCTs involving 773 patients met with our inclusion criteria. In comparison to older fluoroquinolones, azithromycin is marginally better in reducing the chance of CF with RR 0.46 (95% CI 0.25-0.82), while in comparison to ceftriaxone, it significantly reduced the chance of relapse with RR 0.1 (95% CI 0.01- 0.76). There were no serious adverse events reported in any of the trials. Conclusion: Azithromycin can be recommended as a second-line drug in MDR typhoid fever, however, large trials involving pediatric age group patients are recommended to arrive at a definite conclusion.
Keywords: Azithromycin, ceftriaxone, chloramphenicol, enteric fever, fluoroquinolones
Enteric fever caused by Salmonella More Details serotype Typhi (S. Typhi) or Salmonella serotype Paratyphi (S. Paratyphi) remains endemic in many areas of the developing world, causing over 26 million infections and over 200,000 deaths annually.  The incidence is highest in south-central Asia and southeast Asia (over 100/100,000 cases/year), with the highest burden of disease in children aged 2-15 years. 
For decades, chloramphenicol has been highly effective against S. Typhi and S. Paratyphi, and it often remains the antibiotic of choice for the treatment of typhoid fever. However, multidrug-resistant (MDR) strains of S. Typhi and S. Paratyphi (resistant to chloramphenicol, trimethoprim-sulfamethoxazole, and ampicillin) are endemic to many Asian countries. , Expanded-spectrum cephalosporins and fluoroquinolones are used for treating such infections.
Fluoroquinolones have proven to be effective for MDR cases. They are widely regarded as the most effective drug for the treatment of typhoid fever; however, to date, they are restricted from routine use in children, and unfortunately, nalidixic acid-resistant (NaR) isolates of S. Typhi and S. Paratyphi A (defined as susceptible to fluoroquinolone by the microbiology laboratory using the current clinical and laboratory standards institute (CLSI) breakpoints; however, they have reduced susceptibility to fluoroquinolones compared with wild-type strains and also respond less well to fluoroquinolone therapy , ) have also emerged widely.
Ceftriaxone, a third-generation cephalosporin, is highly effective against S. Typhi and has become the standard of care for the treatment of typhoid fever in many parts of the world. However, the cost and parenteral route of administration make ceftriaxone a less-than-ideal treatment alternative. Oral third-generation cephalosporin like cefixime is also widely used, especially in children, however, some studies have shown it to be inferior to other oral agents both in terms of fever clearance time and treatment failure. 
The azalide antimicrobial azithromycin is a further option. Azithromycin, a member of the macrolide class of antibiotics, possesses many characteristics for effective and convenient treatment of typhoid fever, including in vitro activity against many enteric pathogens, excellent penetration into most tissues, and achievement of concentrations in macrophages and neutrophils that are >100-fold higher than concentrations in serum. ,,
However, in in vitro susceptibility testing involving disc diffusion, the appropriate breakpoint recommendations for azithromycin against S. Typhi are still not clear. Patients may respond satisfactorily to azithromycin even if isolates are intermediate according to current guidelines.
This meta-analysis was undertaken to determine the strength of evidence supporting use of azithromycin over the alternate drugs available for treating enteric fever in children and adults.
Studies were identified using electronic database such as MEDLINE and other data at the National Library of Medicine assessed using PUBMED search engine as well as Cochrane Clinical Trial Register using text words and medical subject headings (MeSH) 'enteric fever', 'typhoid fever' 'paratyphoid fever' AND Azithromycin from 1966 to August 2010.
Only randomized control trials (RCTs) comparing azithromycin with other antimicrobials for treatment of uncomplicated enteric fever confirmed with blood or bone marrow or stool culture were included.
Primary outcomes of interest were 'Clinical failure' (CF), defined as persistent symptoms or development of complications requiring prolonged treatment or the addition or change of antimicrobial agent, 'Microbiological failure' (MF) defined as a positive culture from blood, bone marrow, or stool at the end of treatment, and 'Relapse', defined as recurrence of symptoms in addition to a positive culture from blood, bone marrow, or stool within 30 days during the follow-up period.
Secondary outcomes included 'fever clearance time'(FCT), defined as time in hours from the start of the trial or control drug until body temperature falls to values less than 38°C and remains so for a period as specified by trial authors. 'Duration of hospital stay', defined as time in days from entry into trial until discharge, and development of 'adverse drug reactions' (ADR), defined as an injury related to medical management, in contrast to complications of disease.
Each study was assessed individually by two authors (NT and PC). Data were extracted individually from each published manuscript by both the authors and data were only included if the two authors had independently achieved the same results.
Data for dichotomous outcomes, such as clinical failure, microbiological failure and relapse were extracted by recording the total number of participants randomized, those that experienced these outcomes, and the number analyzed.
For continuous outcomes, such as fever clearance time and duration of hospital stay, data were extracted by the total number of participants analyzed, arithmetic means, and standard deviation.
Data on reported adverse events were also extracted.
Methodological quality of the trials was assessed by both the authors based on the method of randomization (generation of allocation sequence), allocation concealment, blinding, and follow-up of participants with culture-proven enteric fever. 
Meta-analysis was conducted using mix software Version 1.7,  with Risk Ratio (RR) for dichotomous data and mean differences (MD) for continuous data, presented with 95% confidence intervals. Random effect model (DerSimonian Laird) was used.
PRISMA guidelines were followed for reporting the study. 
The search 'enteric fever' or 'typhoid fever or 'paratyphoid fever' AND 'azithromycin' with limit 'randomized controlled trials' retrieved seven studies.
All seven studies ,,,,,, including 773 patients met with our inclusion criteria. The methodological quality of all seven included trials was generally high. For instance, all trials used adequate methods to generate the allocation sequence and conceal allocation and all included trials reported well-defined inclusion and exclusion criteria [Table 1].
Criteria for enrolment were patient presenting with signs and symptoms of uncomplicated typhoid fever with positive blood or stool culture for S. Typhi or S. Paratyphi.
All trials excluded pregnant and lactating women, and those with serious underlying diseases, previous antibiotic treatment, severe illness, and history of allergy to any of the study drugs.
All seven trials were described as open for both the participants and physicians.
Dosage of azithromycin used in the studies ranged from 500 mg to 1 g per day (10-20 mg/kg/day) for five to seven days. Two trials treated participants for five days, , whereas the other five trials used a seven-day regimen. ,,,,
Characteristics of the patients
Out of seven, three studies ,, included adults with minimum age of 18 years in two studies , and 15 years in one study,  two studies , included children and adolescents with age range of 3-17 years, and two studies , had mix population ranging from 2-41 years; however, the majority of the population in this study (>70%) were children and adolescents [Table 2].
Mean age of patients in the azithromycin-treated group was 16.4 years and in alternate drug-treated group 16.3 years. Of the study population 55% were male patients while 45% were females.
A total 392 patients were treated with azithromycin while 381 patients were treated with alternate drug. The alternate drug was fluoroquinolone in four studies, ,,, ceftriaxone in two studies , while in one study  it was chloramphenicol [Table 3].
Mean duration of fever before hospitalization was 9.97 and 10.0 days in the alternate drug group and azithromycin-treated group respectively.
All seven trials included participant with MDR strains with percentage of MDR strains ranging from 1.4-90%. In three trials ,, the percentage of MDR strain was more than 50%. Only three trials ,, mentioned presence of NaR strains and in all these three trials the percentage of NaR strain were more than 50%.
Azithromycin versus chloramphenicol:
One study including 77 patients compared azithromycin with chloramphenicol.
Five out of 42 patients in the azithromycin-treated group developed clinical failure compared to five out of 35 in chloramphenicol-treated group with RR of 0.83 (95% CI 0.26 to 2.65). Two out of 35 patients developed MF in the chloramphenicol-treated group, while none of the patients treated with azithromycin developed microbiological failure. However, both these differences were statistically not significant. None of the patients in both the treatment groups developed relapse.
Fever clearance time was marginally less in the azithromycin (mean 98.4 h)-treated group compared to the chloramphenicol-treated group (mean 103.2 h) with MD -4.90 (95% CI -35.0 to 25.28), however, the difference was statistically not significant.
Azithromycin versus fluoroquinolone
Four studies ,,, compared azithromycin with fluoroquinolone. Fluoroquinolones used were gatifloxacin in one study,  ciprofloxacin in one study  and ofloxacin in two studies, , the dose and duration of ofloxacin treatment in both the studies varied [Table 3].
Overall 19 out of 284 patients in the azithromycin-treated group developed clinical failure compared to 35 out of 280 in the fluoroquinolone-treated group with RR of 0.54 (95% CI 0.32 to 0.91) while there was no statistical difference in the rate of microbiological failure or relapse in the two treatment groups with RR 1.0 (95% CI 0.32 to 3.2) and 0.14 (95% CI 0.01 to 1.21) respectively.
There was marked heterogeneity for fever clearance time with no significant difference between the interventions with MD -9.7 (95% CI -34.1 to 14.5).
As nalidixic acid resistance markedly affects response to older fluoroquinolones, we planned to do further analysis to evaluate the effect of NaR status on response to different fluoroquinolones.
Three out of four studies ,, reported about NaR status. All three studies showed high proportion of NaR strains. Out of three, in two studies , the fluoro[quinolone used was ofloxacin while in one  it was gatifloxacin. The proportion of NaR status is mentioned in [Table 2].
Compared to ofloxacin, treatment with azithromycin significantly reduced the clinical failure rate [RR 0.45 (95% CI 0.25 to 0.810] [Figure 1], fever clearance time [MD -27.1 (95% CI -46.3 to -7.87)] as well duration of hospital stay [MD -1.0 (95% CI -1.72 to -0.34)] [Figure 2], while there was no significant difference in microbiological failure or relapse.
While, in a study comparing gatifloxacin with azithromycin, there was no significant difference in clinical failure [RR 1.02(95% CI 0.34 to 3.09)], microbiological failure [RR 1.53 (95% CI 0.26 to 9.03)], relapse [RR 0.20 (95% CI 0.01 to 4.05)] or fever clearance time. There was no data on duration of hospital stay.
Azithromycin versus ceftriaxone
Two studies , compared azithromycin with ceftriaxone. Treatment with azithromycin did not differ significantly from ceftriaxone for clinical failure [RR 2.4 (95% CI 0.49 to 12.2)], or microbiological failure [RR 0.61 (95% CI 0.07 to 4.83)] or in FCT [MD 11.2 (95% CI -4.7 to 27.2)], however, it significantly reduced the chance of relapse. None of the patients in the azithromycin-treated group developed relapse while nine patients out of 66 in the ceftriaxone-treated group relapsed with RR 0.1 (95% CI 0.01 to 0.76) [Figure 3].
Neither trial reported on the duration of hospital stay.
There were no serious adverse events reported in any of the trial.
One patient developed maculopapular rash after first dose of azithromycin. Azithromycin was discontinued and the patient was treated with ceftriaxone.
The most common adverse event reported in both the treatment groups in all the studies was related to the gastrointestinal (GI) tract. Vomiting was most frequent ADR of azithromycin, however it was mild and self-limiting [Table 4].
There were incidents of GI bleeding in four patients of the azithromycin arm of Dolecek 2008, and one event in each arm of Chinh, 2000, however, it was considered part of the typhoid syndrome and the drug was not discontinued [Table 4].
In laboratory investigations, mild increase in liver enzymes was reported by five studies ,,,, and two studies , reported mild thrombocytosis in both the treatment arms, however, it was transient and recovered on treatment discontinuation [Table 4].
Six patients treated with ceftriaxone complained about pain on intramuscular injection, which persisted for 24 h [Table 4].
There is an urgent need to explore the utility and safety of alternate drugs in the treatment of enteric fever due to emergence of multidrug-resistant (MDR) and NaR strains of S. Typhi.
Azithromycin is a potentially useful drug in the treatment of typhoid fever because of its high intracellular tissue penetration and a long elimination half-life (72 h).
This meta-analysis addresses the available evidence on the efficacy and safety of azithromycin in treating enteric fever in comparison to alternate drugs.
Fluoroquinolones were used as alternate drug in four studies while in two studies the alternate drug was ceftriaxone and in one study it was chloramphenicol.
In comparison to fluoroquinolones, azithromycin is marginally better in reducing the chance of CF with RR 0.54 (95% CI 0.32-0.91) however, it was not statistically different than fluoroquinolone in reducing the chance of MF, relapse or in FCT.
The proportion of participants with NaR strains is particularly important for the comparison with fluoroquinolones because such strains may exhibit reduced susceptibility to fluoroquinolones. Chinh 2000,  Dolecek 2008  and Parry 2007  were three of the four trials comparing azithromycin with fluoroquinolones, and all three involved a high proportion of NaR infections.
In comparison with the ofloxacin-treated group, treatment with azithromycin significantly reduced the chance of CF and produced significant reduction in fever clearance time as well as mean duration of hospital stay. However, response of gatifloxacin was comparable to that of azithromycin.
When compared, azithromycin did not differ significantly from ceftriaxone in reducing chance of CF, MF or FCT, however, it significantly reduced the chance of relapse with RR 0.1 (95% CI 0.01 to 0.76). In comparison to ceftriaxone, there was an absolute risk reduction of 13.6% i.e. seven patients need to be treated with azithromycin to prevent one relapse (NNT=7). This appears reasonably good especially when the use of azithromycin was not associated with any serious adverse event and it has the convenience of oral usage. However, it is again to be noted that this result is also based on the analysis of a small number of patients (n=132) from only two studies.
We planned to do subgroup analyses for the pediatric age group, but this was not possible as only two studies involved children and adolescent patients. However, in two other studies that involved a mixed population, more than 70% patients were of the pediatric age group, so overall more than 50% of the study population involved children and adolescents.
All seven trials reported on adverse events. Most adverse events were gastrointestinal in nature, and they were few and mild. Gastrointestinal bleeding occurred in one participant in each of the azithromycin and fluoroquinolone arms in Chinh 2000,  and in four participants in the azithromycin arm in Dolecek 2008.  Laboratory abnormalities like elevation in liver enzymes and platelet counts (thrombocytosis) were also clinically insignificant and self-limiting.
Implications for practice and policy
Evidence from this meta-analysis suggests that azithromycin appears to be as good as the comparator drugs for most outcomes. It appears to be better in comparison to the older fluoroquinolones especially in infections with NaR strains, while in comparison to the newer fluoroquinolones like gatifloxacin, the outcomes were comparable. It was better compared to ceftriaxone, in terms of reducing relapse.
Azithromycin can be recommended as a second-line drug in multidrug-resistant typhoid fever.
However, because of the small number of trials eligible for this meta-analysis and wide confidence intervals for each comparison, we are not able to make firm conclusions as to the benefit of azithromycin over the other drugs.
Large trials involving pediatric patients are needed, especially in outpatient settings to compare azithromycin with other first-line drugs such as oral third-generation cephalosporin.
Moreover, except in Dolecek (2008) trial, which included an 'intention to treat' analysis, in all the other trials, the analysis is based on culture-positive cases. Future trials should analyze their results on an 'intention to treat' basis as well as for the culture-positive cases.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4]