Azithromycin versus ceftriaxone for uncomplicated typhoid fever in children: a systematic review and meta-analysis of randomized controlled trials

Abstract

Background

Typhoid fever caused by Salmonella enterica serovar Typhi remains a major cause of febrile illness among children in low- and middle-income countries, particularly in South and Southeast Asia. The increasing emergence of multidrug-resistant (MDR) and extensively drug-resistant (XDR) strains has substantially limited available treatment options and highlighted the need for effective oral therapies. Azithromycin and ceftriaxone are commonly used for uncomplicated typhoid fever in children; however, evidence comparing their efficacy and safety remains limited.

Objective

To compare the efficacy and safety of azithromycin versus ceftriaxone for the treatment of uncomplicated typhoid fever in children through a systematic review and meta-analysis of randomized controlled trials.

Methods

A comprehensive literature search was conducted in PubMed, the Cochrane Library, ClinicalTrials.gov, Google Scholar, and WHO ICTRP from inception to April 2026 in accordance with PRISMA 2020 guidelines. Randomized controlled trials comparing azithromycin with ceftriaxone in children under 18 years of age with uncomplicated typhoid fever were included. The primary outcome was the clinical cure rate. Secondary outcomes included fever clearance time, microbiological cure, relapse, and adverse events. Data extraction and risk of bias assessment using the Cochrane Risk of Bias 2 tool were performed independently by two reviewers. Meta-analysis was conducted using a random-effects model, and heterogeneity was assessed using the I² statistic.

Results

Six randomized controlled trials involving 492 children (azithromycin = 244; ceftriaxone = 248) were included. All studies were open-label trials conducted in Egypt, India, Pakistan, and Bangladesh. Overall risk of bias was judged to be of some concern, mainly due to the open-label study design. Meta-analysis showed no significant difference between azithromycin and ceftriaxone in clinical cure rate (RR = 0.98, 95% CI 0.93–1.03; P = 0.43) or fever clearance time. Similarly, microbiological cure and adverse event profiles were similar across groups but could not be formally pooled due to heterogeneous reporting. However, azithromycin was associated with fewer relapses than ceftriaxone, though this finding should be interpreted with caution given the low certainty of the evidence. The certainty of evidence was very low for fever clearance and low for the relapse rate. Heterogeneity varied across outcomes, with the highest for fever clearance time.

Conclusion

Azithromycin appears to have comparable efficacy and safety to ceftriaxone for the treatment of uncomplicated typhoid fever in children, with a trend toward fewer relapses that warrants cautious interpretation. Its oral administration may provide a practical alternative to parenteral ceftriaxone in endemic settings. Further large-scale randomized trials with standardized outcome reporting are needed to strengthen the current evidence base.

Clinical trial number

Not applicable.

Systematic review registration

PROSPERO Protocol Number: CRD420261383999.

Link

https://www.crd.york.ac.uk/PROSPERO/view/CRD420261383999.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12879-026-14103-x.

Introduction

Typhoid fever, caused by Salmonella enterica serovar Typhi, continues to be a major infectious disease in low- and middle-income countries, particularly among children [1, 2]. Globally, an estimated 11–21 million cases and more than 100,000 deaths occur each year [1]. The disease is most common in regions with poor sanitation, unsafe water supplies, and limited healthcare facilities, especially in South Asia, Southeast Asia, and sub-Saharan Africa. Children between 2 and 15 years of age are particularly vulnerable, with school-aged children accounting for a large proportion of cases. Countries such as Pakistan, India, and Bangladesh continue to report a high burden of pediatric typhoid fever [2], and contributed most of the studies included in this review.

Children with uncomplicated typhoid fever commonly present with persistent fever, abdominal pain, headache, weakness, and loss of appetite. Diagnosis is usually confirmed through blood culture or serological testing. If treatment is delayed or ineffective, the disease may progress to severe complications such as intestinal perforation, gastrointestinal bleeding, encephalopathy, and sepsis. Therefore, timely and appropriate antibiotic therapy is essential to reduce complications and improve outcomes.

Traditionally, chloramphenicol, ampicillin, and trimethoprim-sulfamethoxazole were widely used for the treatment of typhoid fever. However, the increasing emergence of multidrug-resistant (MDR) Salmonella Typhi strains reduced the effectiveness of these antibiotics and led to greater use of fluoroquinolones and third-generation cephalosporins. More recently, the emergence of extensively drug-resistant (XDR) typhoid, particularly during the outbreak reported in Pakistan between 2016 and 2019, has further limited treatment options [3] by reducing susceptibility to fluoroquinolones and several beta-lactam antibiotics. This changing pattern of resistance has created an urgent need for effective treatment alternatives for children.

Ceftriaxone is commonly used to treat typhoid fever and is considered an effective option, particularly in hospitalized patients [4]. It has reliable activity against Salmonella Typhi and is associated with good clinical response rates. However, because ceftriaxone is administered intravenously or intramuscularly, its use often requires hospital admission or repeated healthcare visits. This can increase treatment costs, prolong hospitalization, and place additional burden on healthcare systems, especially in resource-limited settings.

Azithromycin has emerged as a promising oral alternative for uncomplicated typhoid fever in children [5, 6]. It has several favorable pharmacological properties, including good tissue penetration, high intracellular concentrations, and a long half-life that allows once-daily dosing. Importantly, azithromycin has shown activity against many MDR strains and remains effective against several resistant isolates [7] linked with XDR typhoid. In addition, its oral formulation makes outpatient treatment possible, which may help reduce hospitalization and healthcare costs.

Several randomized controlled trials have compared azithromycin with ceftriaxone in children with uncomplicated typhoid fever [4, 8–12] and evaluated outcomes such as clinical cure, fever clearance time, relapse, microbiological cure, and adverse events. However, differences in study design, sample size, patient population, and antibiotic dosing regimens have resulted in inconsistent findings among individual studies. Therefore, evidence from single studies alone may not provide definitive conclusions regarding the comparative effectiveness of these antibiotics.

This systematic review and meta-analysis were conducted to comprehensively evaluate the available randomized controlled trials comparing azithromycin and ceftriaxone for uncomplicated typhoid fever in children under 18 years of age. By pooling evidence from multiple studies, this review aims to provide clearer evidence regarding the efficacy and safety of these treatment options and support evidence-based management of pediatric typhoid fever in endemic regions.

Although several randomized controlled trials have compared azithromycin and ceftriaxone for the treatment of pediatric typhoid fever, the available evidence remains limited and fragmented across individual studies. Previous systematic reviews have evaluated antibiotic therapy for enteric fever and included comparisons involving azithromycin and ceftriaxone; however, they were not focused exclusively on children with uncomplicated typhoid fever. To our knowledge, no previous systematic review and meta-analysis has specifically synthesized randomized controlled trials comparing azithromycin and ceftriaxone exclusively in pediatric patients with uncomplicated typhoid fever while also evaluating the certainty of evidence using the GRADE framework. Therefore, this review was conducted to provide an updated and clinically focused synthesis of the available evidence and to support evidence-based treatment decisions in pediatric populations.

Methods

This systematic review and meta-analysis were conducted in accordance with PRISMA 2020 guidelines [13]. This review was prospectively registered on PROSPERO (CRD420261383999) on 30th April 2026, prior to commencement of screening and data extraction. No deviations from the registered protocol were made. A comprehensive literature search was conducted on PubMed, the Cochrane Library, ClinicalTrials.gov, and Google Scholar from inception to April 2026. Search terms included: “typhoid fever”, “enteric fever”, “Salmonella typhi,” azithromycin “,” ceftriaxone,” children “, “pediatric”, “randomized controlled trial”. Two reviewers independently screened titles and abstracts, followed by full-text assessment of potentially eligible studies. Any disagreements were resolved through discussion and consensus. Screening was performed using Rayyan. The complete Boolean search strings for each database are provided in Supplementary Table S1. One study could not be included due to full-text unavailability despite attempts to contact the corresponding author.

Data extraction was performed independently by two reviewers. Risk of bias was assessed using the Cochrane Risk of Bias tool (RoB2) [14].

Statistical analysis was performed using RevMan 5.4. Heterogeneity was assessed using the I² statistic [15]. Dichotomous outcomes were analyzed using risk ratio (RR), and Mean Difference (MD) was used for continuous outcomes, both with 95% CI. Because substantial heterogeneity was anticipated (I² >50%), a random-effects model [15] was applied. Publication bias was not formally assessed because fewer than 10 studies were included in the meta-analysis. A sensitivity analysis restricted to blood culture-confirmed studies only (Frenck et al. 2000, Frenck et al. 2004, and Nair et al.) was performed to assess the impact of diagnostic heterogeneity on pooled estimates. The unit of measurement for fever clearance time was verified across all six included studies before pooling the data to ensure consistency and comparability of the outcome measure. For outcomes with zero events in one arm, the default 0.5 continuity correction applied by RevMan’s Mantel-Haenszel method was used.

Eligibility criteria

Studies were included if they were Randomized Controlled Trials (RCTs) involving children aged below 18 with uncomplicated typhoid fever (defined as fever without intestinal perforation, hemorrhage, altered consciousness, or septic shock, per WHO criteria), comparing oral azithromycin with IV or IM ceftriaxone. Studies were excluded if they were non-randomized, included adult patients, enrolled patients with paratyphoid fever (Salmonella Paratyphi), or did not report clinical outcomes.

Outcomes

Clinical cure rate by Day 7 was the primary outcome. Secondary outcomes included fever clearance time and relapse rate. Microbiological cure and adverse events were also recorded, but could not be formally pooled due to heterogeneous reporting across included studies.

Results

A total of 60 studies were identified through database searching. After removing 14 duplicates, 46 studies were screened at the abstract stage. Following abstract screening, 34 studies were excluded. After full-text screening of 12 studies, 6 studies were excluded at this stage, and 6 RCTs met the inclusion criteria and were included in the final meta-analysis (Fig. 1).

Fig. 1.

PRISMA 2020 flow diagram. * Records excluded at abstract screening stage: reasons included non-RCT design, adult population, paratyphoid fever, and absence of relevant comparator. ** Records excluded at full-text review stage: four studies did not meet RCT criteria, one included adult patients, and one had insufficient data for meta-analysis

A total of 6 randomized controlled trials (RCTs) were included in this meta-analysis, comprising 492 participants (AZI: 244; CEF: 248). Study designs included open-label RCTs. Studies were conducted across multiple countries, including Pakistan, Egypt, India, and Bangladesh. The age range of participants was 2–18 years. All studies compared azithromycin (10–20 mg/kg/day orally for 5–7 days) to ceftriaxone (75–100 mg/kg/day IV/IM for 5–10 days). All studies confirmed typhoid using blood culture for Salmonella Typhi, with some also using stool culture, Widal test, or Typhidot (Table 1).

Table 1.

Characteristics of included studies

Study	Year	Country	AZI (n)	CEF (n)	Age Range	Azithromycin Dose & Duration	Ceftriaxone Dose & Duration	Diagnosis Method	Blood Culture Confirmed n/N (%)	Clinical Cure Definition	Study Design
Frenck et al.	2000	Egypt	34	30	4–17 yrs	10 mg/kg/day x 7 days (oral)	75 mg/kg/day x 7 days (IV/IM)	Blood culture (S. typhi)	64/64 (100%)	Resolution of all typhoid-related symptoms or signs by end of Day 7	Open-label RCT
Frenck et al.	2004	Egypt	32	36	3–17 yrs	20 mg/kg/day x 5 days (oral)	75 mg/kg/day x 5 days (IV)	Blood/stool culture (S. typhi)	68/68 (100%)	Resolution of all typhoid-related symptoms within 7 days	Open-label RCT
Nair et al.	2017	India	30	34	3–18 yrs	20 mg/kg/day x 7 days (oral)	75 mg/kg/day x 7 days (IV)	Blood/stool culture (S. typhi)	64/64 (100%)	Becoming afebrile within 7 days of starting treatment	Open-label RCT
Nagaraj et al.	2016	India	63	63	2–12 yrs	20 mg/kg/day x 7 days (oral)	75 mg/kg/day BD x 7 days (IV)	Blood culture +/- Widal test	101/126 (80.2%)	Afebrile by Day 7 (treatment failure = persistence of fever beyond 7 days)	Open-label RCT
Langah et al.	2020	Pakistan	50	50	2–12 yrs	10 mg/kg/day x 7 days (oral)	75 mg/kg/day (divided BD), max 2.5 g/day x 10 days (IV)	Blood culture/Widal/Typhidot	40/100 (40%)	Not defined; Day 7 cure rate not reported	Open-label RCT
Jahan et al.	2019	Bangladesh	35	35	2–12 yrs	20 mg/kg/day (2 div. doses) x 7 days (oral)	100 mg/kg/day x 10 days (IV)	Blood culture +/- Widal test	54/70 (77.1%)	Not defined; Day 7 cure rate not reported	Open-label RCT

AZI= Azithromycin; CEF= Ceftriaxone; BD= twice daily; IV= intravenous; IM= intramuscular; RCT= Randomized Controlled Trial. *Langah et al. and Jahan et al. were excluded from the clinical cure meta-analysis as they did not report the Day 7 cure rate

Quality assessment

Risk of bias was assessed using the Cochrane Risk of Bias 2 (RoB2) tool [14] across five domains: bias arising from the randomization process, deviations from intended interventions, missing outcome data, measurement of outcomes, and selection of reported results (Supplementary Table S2). Four of the six trials (Frenck 2000, Frenck 2004, Nair 2017) described an adequately generated and concealed randomization sequence, while three trials (Nagaraj 2016, Langah 2020, Jahan 2019) reported randomization only as a “lottery” or “lots” method without describing sequence generation or concealment, raising some concerns on this domain. All six trials were open-label, as blinding was not feasible given the differing routes of administration (oral azithromycin versus parenteral ceftriaxone); applying the RoB2 signaling questions for Domain 2, five trials showed no consequential deviations from the intended intervention and used an appropriate intention-to-treat-style analysis, and were judged low risk on this domain despite being open-label. This is most relevant to subjective, clinician-assessed outcomes — clinical cure and fever clearance time — and less relevant to laboratory-confirmed outcomes such as microbiological cure and culture-confirmed relapse. Nagaraj 2016 was judged at high overall risk of bias, driven by an unclear intention-to-treat analysis following the crossover of three azithromycin-arm treatment failures to ceftriaxone, and because its reported duration-of-hospital-stay outcome is structurally confounded by the route of administration rather than reflecting comparative drug efficacy. None of the six trials reported prospective registration in a public trial’s registry, so selective outcome reporting could not be fully excluded for any study.

The overall risk of bias assessment is presented in Fig. 2.

Fig. 2.

Risk of bias assessment of included studies

A pooled analysis of four randomized controlled trials (Frenck et al., Frenck et al. 2004, Nagaraj et al., and Nair et al.) [8–11], involving 322 children (159 in the azithromycin group and 163 in the ceftriaxone group) assessed the clinical cure rate. The pooled analysis showed no significant difference between azithromycin and ceftriaxone in achieving clinical cure (RR 0.98, 95% CI 0.93–1.03; P = 0.43). No heterogeneity was observed among the included studies (I² =0%), indicating consistent results across trials (Fig. 3).

Fig. 3.

Forest plot of clinical cure. Events = number of children achieving clinical cure (Fig. 3) / remaining relapse-free (Fig. 6). Total (95% CI) reports the pooled effect estimates; Total events reports the raw cumulative event count. M-H = Mantel-Haenszel method, used for pooling dichotomous outcomes under the random-effects model

Secondary outcomes

Six studies (Frenck et al., Frenck et al. 2004, Jahan et al., Langah et al., Nagaraj et al., and Nair et al.) [4, 8–12], including 492 children (244 receiving azithromycin and 248 receiving ceftriaxone), reported fever clearance time. The pooled mean difference was − 0.11 days (95% CI − 1.41 to 1.18), showing no significant difference between the two treatment groups (P = 0.86). Given the very high heterogeneity (I² = 98%), this pooled estimate should be interpreted as illustrative only, rather than as a precise measure of effect.

However, substantial heterogeneity was present among the studies (I² =98%, P < 0.0001). This variation may be due to differences in disease severity, treatment regimens, and patient characteristics across studies (Fig. 4). All six studies reported fever clearance time in days, and no unit inconsistency was identified across studies. The substantial heterogeneity observed (I² = 98%) may be partly attributable to the study by Jahan et al., in which the ceftriaxone group exhibited a markedly longer mean fever clearance time (7.14 ± 0.42 days) compared with the other included studies (range: 3.6–4.5 days). This discrepancy may reflect differences in patient characteristics, disease severity, treatment protocols, or study settings, thereby contributing to the observed between-study heterogeneity.

Fig. 4.

Forest plot of fever clearance time. IV = inverse Variance method, used for pooling continuous outcomes under the random-effects model. Mean and SD are expressed in days

Subgroup analysis

A subgroup analysis according to the azithromycin dosing regimen was performed to assess the source of heterogeneity observed in fever clearance time. In the 10 mg/kg/day subgroup, including studies by Frenck et al. and Langah et al. (n = 164), no significant difference was found between groups (MD 0.08 days, 95% CI − 0.22 to 0.39; P = 0.59). No heterogeneity was observed within this subgroup (I²=0%, Chi²=0.30, P = 0.58).

In the 20 mg/kg/day subgroup, including studies by Frenck et al. 2004, Jahan et al., Nagaraj et al., and Nair et al. (n = 328), the pooled analysis also showed no significant difference (MD − 0.22 days, 95% CI − 2.07 to 1.63; P = 0.82). However, substantial heterogeneity remained within this subgroup (I²=97%, Chi²=116.62, P < 0.00001).

The test for subgroup differences was not statistically significant (Chi²=0.10, df = 1, P = 0.75, I²=0%), suggesting that the azithromycin dosing regimen was unlikely to account for the overall heterogeneity. The residual heterogeneity in the 20 mg/kg/day subgroup may be related to variations in baseline disease severity, geographic setting, and ceftriaxone dosing among the included studies (Fig. 5).

Fig. 5.

Subgroup analysis of fever clearance time by azithromycin dosing regimen (10 mg/kg/day vs. 20 mg/kg/day). IV = inverse Variance method, used for pooling continuous outcomes under the random-effects model. Mean and SD are expressed in days

Sensitivity analysis

A leave-one-out sensitivity analysis was carried out for fever clearance time to investigate the source of the high heterogeneity observed in the overall analysis (I²=98%). Each study was removed one at a time, and the pooled results were recalculated to evaluate its impact on the findings.

Among all included studies, exclusion of Jahan et al. resulted in a noticeable reduction in heterogeneity, although the pooled effect remained statistically non-significant (MD 0.26 days, 95% CI − 0.17 to 0.69; P = 0.23). In comparison, exclusion of the other studies, including Frenck et al. 2000, Frenck et al. 2004, Langah et al., Nagaraj et al., and Nair et al., did not substantially affect either the pooled estimate or the degree of heterogeneity, which remained consistently high.

These results indicate that Jahan et al. was the major contributor to the observed heterogeneity. One possible explanation is the relatively longer fever clearance time reported in the ceftriaxone group of this study compared with the other included trials, which may reflect differences in patient severity, study setting, or treatment regimen. Overall, the sensitivity analysis suggests that the main findings of the meta-analysis remained stable and were not significantly influenced by the exclusion of any single study (Table 2).

Table 2.

Sensitivity analysis- effect of individual study exclusion on fever clearance

Study Excluded	MD (95% CI)	I²	P value
None (baseline — all studies)	−0.11 [− 1.41, 1.18]	98%	0.86
Jahan et al. 2019	0.26 [− 0.17, 0.69]	58%	0.23
Frenck et al. 2000	−0.17 [− 1.64, 1.29]	98%	0.82
Frenck et al. 2004	−0.31 [− 1.69, 1.08]	98%	0.66
Langah et al. 2020	−0.14 [− 1.69, 1.42]	98%	0.86
Nagaraj et al. 2016	−0.05 [− 1.54, 1.43]	98%	0.94
Nair et al. 2017	−0.33 [− 1.71, 1.05]	98%	0.64

MD=Mean Difference; CI = Confidence interval; I²= Inconsistency statistic

Relapse rate was reported in four studies by Frenck et al., Nagaraj et al., and Nair et al., [8–11] involving a total of 322 children (azithromycin = 159, ceftriaxone = 163).

No relapse was observed in any of the patients who received azithromycin, whereas 17 cases of relapse were reported in the ceftriaxone group. When the data were pooled, azithromycin was associated with fewer relapses compared with ceftriaxone (RR 0.11, 95% CI 0.03–0.48; P = 0.003). However, these findings should be interpreted cautiously, given the low certainty of evidence, small number of events, and variation in relapse definitions and follow-up windows across studies.

The results were highly consistent across all studies, with no evidence of heterogeneity (I²=0%, Chi²=0.19, df = 3, P = 0.98). In the forest plot, all individual studies favored azithromycin, and the overall pooled estimate also clearly supported azithromycin, with the confidence interval entirely on its side of benefit (Fig. 6). Relapse definitions and follow-up windows across contributing studies are summarized in Table 3.

Fig. 6.

Forest plot of relapse rate. Events = number of children remaining relapse-free. Total (95% CI) reports the pooled effect estimate and confidence interval; Total events reports the raw cumulative event count across all included studies. M-H = Mantel-Haenszel method, used for pooling dichotomous outcomes under the random-effects model

Table 3.

Relapse definition and follow-up window across contributing studies

Study	Relapse Definition	Follow-up Window	Relapse Events (AZI/CEF)
Frenck et al. 2000	Recurrence of fever with signs or symptoms of typhoid fever along with isolation of S. typhi or S. paratyphi from the blood	4 weeks (28 days)	0/34 vs. 6/30
Frenck et al. 2004	Recurrence of fever and clinical features of typhoid within 30 days of completing therapy, along with isolation of S. typhi from the blood	30 days	0/32 vs. 6/36
Nair et al. 2017	Recurrence of typhoid-related symptoms with isolation of S. typhi from blood specimens	30 days (1 month)	0/30 vs. 5/34
Nagaraj et al. 2016	Recurrence of fever investigated for typhoid fever with positive stool culture	30 days	0/63 vs. 2/63

AZI = Azithromycin; CEF = Ceftriaxone. Note: Follow-up windows varied slightly across studies (28–30 days), which may have contributed to differences in relapse detection. All relapse events occurred in the ceftriaxone group; no events were recorded in the azithromycin group across all four studies

Although six studies were included in the overall meta-analysis, relapse data were available in only four trials. The remaining two studies did not report relapse outcomes and, therefore, could not be included in this specific analysis. These results are summarized in Table 4.

Table 4.

Summary of outcomes

Outcome	Studies (n)	Contributing Studies	Participants	Effect Estimate (95% CI)	I2	P value	Favours
Clinical Cure Rate	4	Frenck 2000, Frenck 2004, Nair et al., Nagaraj et al.	322	RR 0.98 [0.93, 1.03]	0%	0.43	Neither
Fever Clearance Time	6	Frenck 2000, Frenck 2004, Nair et al., Nagaraj et al., Langah et al., Jahan et al.	492	MD -0.11 [-1.41, 1.18]	98%	0.86	Neither
Relapse Rate	4	Frenck 2000, Frenck 2004, Nair et al., Nagaraj et al.	322	RR 0.11 [0.03, 0.48]	0%	0.003	Azithromycin
Microbiological Cure	-	Could not be formally pooled due to heterogeneous reporting	-	Not pooled	-	-	-
Adverse Events	-	Could not be formally pooled due to heterogeneous reporting	-	Not pooled	-	-	-

RR= Risk Ratio; MD= Mean Difference; CI= Confidence Interval; I2 = Inconsistency statistic; - = not applicable (outcome could not be formally pooled

A sensitivity analysis restricted to studies using blood culture confirmation only (Frenck et al. 2000, Frenck et al. 2004, Nair et al.) showed clinical cure RR 0.96 (95% CI 0.90–1.02; I²=0%), fever clearance MD favoring ceftriaxone (I²=42%; P = 0.03), and relapse RR 0.10 (95% CI 0.02–0.50; I²=0%; P = 0.005), suggesting diagnostic heterogeneity contributed substantially to the high I² observed in the primary analyses.

Microbiological cure

Microbiological cure — defined as a sterile blood culture on follow-up — was reported by four of the six included studies. Frenck et al. 2000 reported cure in 33 of 34 (97%) azithromycin recipients versus 29 of 30 (97%) ceftriaxone recipients. Frenck et al. 2004 reported cure in 32 of 32 (100%) azithromycin recipients versus 35 of 36 (97%) ceftriaxone recipients. Nair et al. reported cure in all azithromycin recipients evaluated versus 33 of 34 (97%) ceftriaxone recipients. Jahan et al. reported that repeat blood culture was negative in all 54 previously culture-positive participants in both arms. Nagaraj et al. and Langah et al. did not report microbiological cure as a discrete outcome. Across all reporting studies, cure rates exceeded 95% in both arms with no apparent between-group difference; however, formal pooling was not undertaken given the differing follow-up culture timepoints (day 7, 8, or 10) across studies and the absence of this outcome in two trials.

Adverse events

No serious adverse events were reported in any of the six included trials. Among the three studies reporting adverse event data, mild, self-limited gastrointestinal symptoms — principally vomiting, diarrhea, nausea, abdominal pain, and anorexia — were the most common adverse events in both treatment arms, with no statistically significant between-group differences in Frenck et al. 2004 or Nair et al. Frenck et al. 2000 additionally reported that injection-site pain was the most common adverse event in the ceftriaxone group, while vomiting was more frequent with azithromycin; asymptomatic thrombocytosis and mild transaminase elevations occurred at similar rates in both arms. Nagaraj et al., Langah et al., and Jahan et al. did not report adverse event data. Given that half of the included trials did not assess this outcome, and that reporting formats varied across the remaining studies, adverse events were not formally pooled.

Discussion

Azithromycin showed similar clinical efficacy to ceftriaxone for the treatment of uncomplicated typhoid fever in children. Across six randomized controlled trials involving 492 pediatric patients [4, 8–12], there was no significant difference between the two antibiotics in terms of clinical cure rate or fever clearance time. These findings suggest that oral azithromycin may serve as an effective alternative to parenteral ceftriaxone in clinically stable children with uncomplicated disease.

The findings of this review are consistent with previous studies. Frenck et al. (2000, 2004) first reported that azithromycin achieved cure rates comparable to ceftriaxone in Egyptian children with culture-confirmed typhoid fever [8, 9]. Similar results were later observed in studies from India by Nair et al. and Nagaraj et al., [10, 11], which demonstrated comparable clinical and microbiological responses between the two treatments. Langah et al. also reported favorable outcomes with azithromycin in Pakistani children [4], an important observation given the increasing burden of drug-resistant typhoid in the region [3]. Comparable findings were further supported by Jahan et al. in Bangladesh [12]. By combining data from all available randomized trials, the present meta-analysis provides stronger pooled evidence and improves the precision of treatment estimates compared with individual studies alone.

The effectiveness of azithromycin against Salmonella Typhi is biologically plausible [5, 6]. Salmonella Typhi is an intracellular pathogen that survives within macrophages of the reticuloendothelial system. Azithromycin achieves high intracellular concentrations and has a prolonged tissue half-life [6], allowing sustained activity against intracellular bacteria. These pharmacokinetic properties may explain why oral azithromycin produced outcomes comparable to intravenous ceftriaxone despite differences in route of administration.

Antibiotic stewardship implications

The findings of this review have important implications for antimicrobial stewardship, particularly in typhoid-endemic and resource-limited settings [16]. Use of oral azithromycin can reduce the need for intravenous or intramuscular therapy, thereby avoiding complications related to venous access such as phlebitis, catheter-associated infections, and injection-site pain. This is especially relevant in children, where maintaining venous access may be challenging. Alongside antibiotic therapy, typhoid conjugate vaccines (TCVs) represent a critical prevention strategy. WHO-prequalified TCVs are being deployed in high-burden countries, providing an essential complement to treatment-based approaches in reducing the overall typhoid disease burden [17].

In addition, oral therapy allows outpatient management or earlier hospital discharge, which may help reduce healthcare costs, decrease hospital overcrowding, and lower the risk of hospital-acquired infections [2]. In many low-resource settings, ceftriaxone administration requires additional equipment, nursing support, and monitoring, all of which increase treatment costs. From a stewardship perspective, reducing unnecessary use of broad-spectrum parenteral antibiotics may also help limit antimicrobial resistance. Based on the available evidence, azithromycin may be considered an appropriate first-line option for children with uncomplicated typhoid fever who are able to tolerate oral medication and reside in areas where local susceptibility patterns support its use. However, increasing reports of azithromycin resistance among Salmonella Typhi strains highlight the need for continuous antimicrobial surveillance and resistance-guided treatment decisions. Ceftriaxone may remain preferable in patients with severe disease, inability to tolerate oral therapy, or in settings where azithromycin resistance is documented or highly prevalent.

Strengths

This meta-analysis has several strengths. It included six randomized controlled trials conducted across multiple countries, including Egypt, India, Pakistan, and Bangladesh, improving the overall applicability of the findings to typhoid-endemic regions. A systematic search strategy was performed according to PRISMA 2020 guidelines [13] to ensure comprehensive study identification and transparent reporting. Risk of bias was assessed using the Cochrane RoB2 tool [14], and statistical analyses were conducted using RevMan 5.4 with appropriate assessment of heterogeneity. In addition, the certainty of evidence for each outcome was evaluated using the GRADE approach [18].

Limitations

Several limitations should be acknowledged. First, all included trials were open-label studies, which may have introduced performance bias, particularly for subjective outcomes such as fever clearance time. However, more objective outcomes, including microbiological cure, are less likely to have been substantially affected by the absence of blinding. Second, there was variation in azithromycin dosing regimens across studies, ranging from 10 mg/kg/day to 20 mg/kg/day for five to seven days, which may have influenced treatment outcomes and limited direct comparison between trials; the optimal dosing regimen for paediatric typhoid fever therefore remains uncertain. Third, all eligible RCTs were identified from South Asia and Egypt, which represent the highest-burden regions for typhoid fever. While this reflects the geographic distribution of the evidence base rather than a search restriction, it may limit the generalizability of findings to other endemic regions with different antimicrobial resistance patterns, healthcare infrastructures, and transmission dynamics. Fourth, none of the included trials specifically evaluated confirmed extensively drug-resistant (XDR) typhoid fever, and antimicrobial susceptibility testing was not systematically reported. Given the growing emergence of azithromycin-resistant Salmonella Typhi, particularly among XDR strains in Pakistan and Bangladesh, the findings of this review may not be directly applicable to these settings, and further research is needed.

In addition, diagnostic heterogeneity across the included studies represents a meaningful limitation. The diagnosis of typhoid fever was established using different methods: blood culture confirmation was used as the sole criterion in three trials (Frenck et al. 2000 and 2004; Nair et al. 2017), while the remaining three trials also accepted positive Widal serology or Typhidot as alternative diagnostic criteria (Nagaraj et al. 2016; Jahan et al. 2019; Langah et al. 2020). Importantly, the proportion of participants with blood culture confirmation varied substantially among these latter trials, approximately 80% in Nagaraj et al., approximately 77% in both arms of Jahan et al., but fewer than 50% in Langah et al., meaning that a proportion of enrolled children in Langah et al. may not have had confirmed Salmonella Typhi infection. Both the Widal test and Typhidot have limited sensitivity and specificity in endemic settings and may yield false-positive results, particularly in the presence of non-typhoidal Salmonella. This diagnostic variability may have contributed to heterogeneity in observed clinical outcomes and has been reflected in the GRADE downgrading decisions under the indirectness domain; pooled estimates should therefore be interpreted with appropriate caution. A sensitivity analysis restricted to the three trials using blood culture as the sole diagnostic criterion was performed; pooled estimates from this subset should still be interpreted with caution given the smaller sample. Further subgroup analyses stratified by ceftriaxone dose, diagnostic method, disease severity, age, care setting, or duration of fever before enrolment were precluded by the small number of included trials, which would have resulted in strata containing only one or two studies each and therefore unreliable pooled estimates.

Finally, publication bias could not be formally assessed using funnel plots because fewer than ten studies were included; this represents an inherent methodological limitation. Data on length of hospital stay and duration of patient follow-up were not consistently reported across included studies, precluding formal analysis of these clinically relevant outcomes. Conference abstracts and preprint servers were not searched, as typhoid fever RCTs in children are rarely disseminated through these channels, but unpublished negative trials cannot be excluded.

Certainty of evidence (GRADE)

The certainty of evidence was evaluated using the GRADE approach (Table 5). The evidence for clinical cure at Day 7 was rated moderate certainty (RR 0.98, 95% CI 0.93–1.03), downgraded by one level because all included studies used an open-label design, which increased the risk of performance and detection bias. For fever clearance time, the evidence was rated very low certainty (MD − 0.11 days, 95% CI − 1.41 to 1.18), downgraded for risk of bias, very serious inconsistency (I²=98%, P < 0.00001), and imprecision. The extreme heterogeneity was primarily driven by Jahan et al., in which the ceftriaxone arm showed markedly prolonged defervescence (7.14 days) relative to all other studies (3.6–5.5 days); sensitivity analysis confirmed that exclusion of this study reduced I² to 58%. For relapse at Day 30, the evidence was rated low certainty (RR 0.11, 95% CI 0.03–0.48), downgraded for risk of bias and imprecision. No relapses occurred in any azithromycin-treated patient (0/159, 0.0%), while 17 of 163 ceftriaxone-treated patients relapsed (10.4%). This finding should be interpreted cautiously, given the small number of events and low certainty of evidence (Table 5).

Table 5.

GRADE summary of findings: azithromycin versus ceftriaxone for uncomplicated typhoid fever in children

Certainty assessment							Nᵒ of patients		Effect		Certainty	Importance
No of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Azithromycin	Ceftriaxone	Relative (95% CI)	Absolute (95% CI)
Clinical Cure (follow-up: mean 7 days; assessed with: Blood Culture/ Clinical assessment)
4	randomised trials	seriousb	not serious	not serious	not serious	none	149/159 (93.7%)	156/163 (95.7%)	RR 0.98 (0.93 to 1.03)	19 fewer per 1,000 (from 67 fewer to 29 more)	⊕⊕⊕◯ Moderatea	CRITICAL
								0.0%		0 fewer per 1,000 (from 0 fewer to 0 fewer)
								0.0%		0 fewer per 1,000 (from 0 fewer to 0 fewer)
Fever clearance (follow-up: mean 7 days; assessed with: Clinical Assessment(days)
6	randomised trials	seriousb	very seriousc	not serious	seriousd	none	244	248	-	MD 0.11 days lower (1.41 lower to 1.18 higher)	⊕◯◯◯ Very lowb, c,d	CRITICAL
Relapse Rate (follow-up: mean 30 days; assessed with: Clinical follow-up)
4	randomised trials	seriouse	not serious	not serious	seriousf	none	0/159 (0.0%)	17/163 (10.4%)	RR 0.11 (0.03 to 0.48)	93 fewer per 1,000 (from 101 fewer to 54 fewer)	⊕⊕◯◯ Lowe, f	CRITICAL
								0.0%		0 fewer per 1,000 (from 0 fewer to 0 fewer)

CI: confidence interval; MD: mean difference; RR: risk ratio

Explanations

a. All included studies were open-label RCTs, introducing performance and detection bias

b. All included studies were open-label RCTs, introducing performance and detection bias

c. Very serious inconsistency due to substantial unexplained heterogeneity (I²=98%)

d. Wide inconsistency interval crossing the line of no effect (MD 0.11 [-1.41, 1.18]), indicating imprecision in the estimate

e. All included studies were open-label RCTs, introducing performance and detection bias

f. Downgraded for serious imprecision due to the small number of relapse events (17 in total ) and the wide confidence interval (RR 0.11, 95% CI 0.03–0.48)

Conclusion

This systematic review and meta-analysis suggest that azithromycin may have clinical efficacy similar to ceftriaxone for the treatment of uncomplicated typhoid fever in children, based on a limited number of open-label randomized controlled trials. No significant differences were observed in clinical cure rate or fever clearance time between treatment groups. Azithromycin was associated with fewer relapses in the available studies; however, this finding should be interpreted cautiously because relapse events were limited and the certainty of evidence was low. Given its oral administration and practical advantages, azithromycin may be a useful option for clinically stable children in settings where local antimicrobial susceptibility patterns support its use. Nevertheless, concerns regarding emerging antimicrobial resistance highlight the importance of antibiotic stewardship and continued surveillance. Future well-designed multicenter randomized controlled trials with standardized outcome definitions, microbiological confirmation, resistance profiling, and consistent follow-up are needed to strengthen the evidence base.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Funding

The authors received no specific funding for this work.

Data availability

All data generated or analyzed during this study are included in this published article.

Declarations

Ethical approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.