From the systematic search conducted, 8261 articles were identified, and after removing duplicates, 5717 were screened for title and abstract. 77 articles were deemed eligible for review, of which 27 studies were included after a full-text review: 9 systematic reviews [29,30,31,32,33,34,35,36,37] (including 6 with meta-analyses), 4 randomized controlled trials [38,39,40,41], and 14 observational studies [42,43,44,45,46,47,48,49,50,51,52,53,54,55] (including 11 cohort studies and 3 Before and After studies) were finally included (Fig. 1). The list of excluded articles and the reason for exclusion are reported in the supplementary material (Supplement, Sects. 3 and 4). For the review by Lodha et al. [31], which included 29 articles, only the 7 articles conducted in HIC countries were considered available for the current study, whereas those conducted in LMIC were excluded.
PRISMA flowchart of included studies. Legend: n, number
The characteristics of the studies included, and the main outcomes are reported in Table 1.
Overall, the studies were considered of low to moderate quality (Supplementary Material, Sects. 5, 6 and 7). Two reviews [36, 37] were subsequently excluded due to their low quality. The quality of evidence evaluated using the GRADE method was generally very low or low, only occasionally moderate (Supplementary Material, Sect. 8). The recommendations with the certainty of evidence by GRADE and the strength of each recommendation formulated considering both the evidence and the expert opinion panel members is reported in Table 2.
The following results and recommendations are mainly based on the outcome of “treatment failure” because a paucity of studies was found regarding the risk of developing resistance to antibiotics and the risk of adverse events associated with antibiotic therapy without significant results.
Question 1. What is the first-line antibiotic for treating mild-to-moderate CAP in a child under five years old with a complete immunization schedule (at least two doses of hexavalent and pneumococcal vaccines)?
Recommendation 1
In children under five with a complete immunization schedule and mild-moderate CAP, amoxicillin prescription is recommended. (Very low quality of evidence for amoxicillin vs. broad-spectrum antibiotics, amoxicillin vs. macrolides, beta-lactams vs. beta-lactams plus macrolides, aminopenicillin vs. broad-spectrum antibiotics; low quality of evidence for beta-lactams vs. macrolides. Strong recommendation in favor of the intervention. 100% panel consensus).
From the analysis of the collected scientific evidence (Table 1), it has been suggested that therapy with narrow-spectrum antibiotics such as amoxicillin and intravenous aminopenicillins is not, in most cases, inferior in terms of treatment failure risk compared to therapy with broad-spectrum antibiotics (such as amoxicillin-clavulanate or 2nd or 3rd generation cephalosporins) or combination therapy of beta-lactams with macrolides [42,43,44,45, 48, 49, 51,52,53] (very low quality of evidence for amoxicillin vs. broad-spectrum antibiotics, amoxicillin vs. macrolides, beta-lactams vs. beta-lactams plus macrolides, aminopenicillin vs. broad-spectrum antibiotics; low quality of evidence for beta-lactams vs. macrolides).
In the study published by Ambroggio et al. in 2015 on a cohort of 1164 pediatric patients with pneumonia treated as outpatient, no statistically significant difference in treatment failure at 7 and 14 days was found between the group treated with beta-lactams (penicillins, aminopenicillins and third generation cephalosporins) and that treated with macrolides [Adjusted Odds Ratio (AOR): 0.90; 95% Confidence Interval (CI): 0.37–2.22] [42]. The same result was observed comparing children treated with beta-lactams to those treated with the combination therapy of beta-lactams and macrolides (cohort of 717 children, treatment failure at 7 days OR 1.33, 95%CI 0.74–2.39 and at 14 days OR 1.34, 95%CI 0.83–2.18) [44]. Instead, the study published by Lipsett et al. in 2021, on a cohort of 252,177 children with non-severe community-acquired pneumonia treated as outpatient, showed slightly different results, as combination therapy with beta-lactams (both narrow and broad spectrum) plus macrolides showed a reduction in the number of new antibiotics prescriptions compared to beta-lactams alone [narrow-spectrum plus a macrolide vs. narrow-spectrum (OR 0.47, 95% CI: 0.39–0.57); broad-spectrum plus a macrolide vs. broad-spectrum (OR 0.48, 95% CI: 0.41–0.56)]. On the contrary, an increase in the number of new antibiotic prescriptions was observed comparing the group treated with broad-spectrum antibiotics compared to the narrow-spectrum one (OR 1.15; 95% CI: 1.09–1.21), whereas no difference was observed comparing children treated with narrow-spectrum and those treated with macrolides (OR 1.15; 95% CI: 1.09–1.21) [46]. Nevertheless, this study included children aged 1 month to 18 years without stratifying the analysis by age group. This may have introduced a bias, as older children could benefit more from macrolide therapy for atypical pneumonia.
Different systematic reviews report on the treatment of microbiologically confirmed pneumonia caused by M. pneumoniae. No statistically significant differences were observed in patients treated with macrolides compared to those treated with beta-lactams [29, 30].
The systematic review with meta-analysis by Biondi et al., published in 2015, included 16 articles, published between 1967 and 2012, regarding the treatment of microbiologically confirmed pneumonia caused by Mycoplasma pneumoniae, without age distinction. Only five studies were included in the meta-analysis, showing a risk difference of 0.12 (95%CI, -0.04 to 0.20). Indeed, macrolides did not show a significant clinical benefit compared to beta-lactams.
Nevertheless, the studies included in the reviews presented significant heterogeneity, with the inclusion of a population with both lower and upper respiratory tract infections One of the studies under consideration documented a decrease in treatment failure risk among subjects administered macrolides [46]. However, these findings pertained to the broader pediatric population up to 18 years old, without specific disaggregation for those under five. This detail is pivotal as it’s recognized that macrolide treatment might be beneficial for older children, who are epidemiologically more prone to atypical pneumonia.
Indeed, atypical pneumonia may resolve spontaneously or benefit from macrolide therapy to support recovery. Nevertheless, while the authors aimed to balance comparison groups regarding pneumonia severity, it’s possible that children receiving macrolides had less severe pneumonia initially. It’s also important to recognize that a significant proportion of pediatric pneumonia cases are viral. Therefore, the increased efficacy observed in some groups of children treated with macrolides may be influenced more by the underlying causes than just the antibacterial effects.
The recommendation to use amoxicillin, albeit based on low-quality studies due to the retrospective observational nature of the studies and the diagnostic uncertainty that characterised all these studies, confirms what has been reported by the main guidelines published before 2012 by IDSA [8] the British Thoracic Society [9] and by a discussion paper published in collaboration with ESPID [10]. The guidelines rely on pre-2012 studies, excluded from this review due to agreed temporal constraints on article selection. The efficacy of amoxicillin in treating Streptococcus pneumoniae infections has long been established and substantiated scientifically. Amoxicillin stands as a preferred therapeutic option for bacterial infections caused by this pathogen, encompassing lung infections, otitis, and sinusitis, owing to its demonstrated efficacy, minimal side effects, and affordability [56,57,58,59]. This historical evidence, coupled with expert consensus, has enabled the formation of robust recommendations, even in light of limited or very low-quality evidence.
Question 2. What is the first-line antibiotic for treating mild-moderate CAP in a child over five years old with a complete immunization schedule (at least two doses of hexavalent and pneumococcal vaccines)?
Recommendation 2
In children over five years of age with a complete immunization schedule and mild-moderate CAP, treatment with amoxicillin is recommended. (Very low quality of evidence for amoxicillin vs. broad-spectrum antibiotics, aminopenicillin vs. broad-spectrum antibiotics. Strong recommendation in favor of the intervention. 100% panel consensus).
In children who remain well-appearing and do not require hospitalization, macrolide therapy is recommended if there is no clinical improvement after 48 h of amoxicillin treatment. (Very low quality of evidence amoxicillin vs. macrolides, beta-lactams vs. beta-lactams pus macrolides; low quality of evidence for beta-lactams vs. macrolides. Strong recommendation in favor of the intervention. 100% panel consensus).
For children older than 5 years, it appears that combination therapy involving beta-lactams alongside macrolides may mitigate the risk of treatment failure compared to beta-lactam monotherapy [42, 44, 46], although other studies showed no difference in treatment failure [29, 30, 45, 48, 49, 51, 53] (Table 1). However, the evidence quality remains low due to the retrospective nature of most of the studies included, the uncertainty of the diagnosis and the considerable heterogeneity among study outcomes. Notably, one study documented a statistically significant reduction in hospitalization duration in moderate CAP cases treated with combination therapy, particularly in those over 6 years old, although no difference was noted in rehospitalization risk [43]. Conversely, another study failed to observe any significant difference in clinical improvement between children administered combination therapy versus monotherapy, prompting queries regarding the practical efficacy of macrolide antibiotic regimens, especially in hospitalized patients [52]. This observation is significant, given that atypical bacterial pneumonia cases often do not necessitate hospitalization and can resolve spontaneously. However, initiating specific antibiotic therapy may expedite recovery, potentially leading to a quicker resolution [52].
Given these results, it is recommended to consider starting macrolide in children over 5 years with CAP who have not shown clinical improvement after 48 h of amoxicillin treatment but remain well-appearing and do not require hospitalization.
This aligns with the IDSA guidelines [8], which recommend considering this therapy for children over 5 years old. Considering the growing prevalence of macrolide-resistant S. pneumoniae strains and the substantial rate of spontaneous eradication of Mycoplasma pneumoniae documented by certain studies before 2012, employing macrolides as first-line therapy is not deemed appropriate [60,61,62].
Question 3. What is the first line antibiotic for treating mild-moderate CAP in a child without a complete immunization schedule (< 2 doses of hexavalent and pneumococcal vaccines)?
Recommendation 3
In children who are either unimmunized or have incomplete immunization against S. pneumoniaebut are immunized for H. influenzaetype b (having received at least two doses of hexavalent vaccine but less than two doses of pneumococcal vaccine) and present with mild-moderate CAP, monotherapy with amoxicillin is recommended as the first-line therapy. (Low-quality evidence. Strong recommendation in favor of the intervention. Panel consensus 100%)
For children who are either unimmunized or have incomplete immunization coverage for both H. influenzaetype b and S. pneumoniae(having received less than two doses of hexavalent and pneumococcal vaccines), first-line therapy with amoxicillin-clavulanate or second or third-generation cephalosporins is recommended. (Low-quality evidence. Weak recommendation in favor of the intervention. Panel consensus 100%)
Literature evidence concerning first-line therapy in unimmunized children is scarce, particularly in outpatient settings (Table 1). Studies focusing on unimmunized children primarily involve those hospitalized for non-severe pneumonia, often immunized for H. influenzae type b but not S. pneumoniae. Among the available studies, there’s an indication that oral narrow-spectrum antibiotic therapy is comparable in efficacy to intravenous therapy [50, 54, 63].
Regarding macrolide therapy, studies by Roord and Block (included in the meta-analysis published by Lodha [31]) suggested the effectiveness of these antibiotics in treating CAP in unimmunized children, considering the high prevalence of isolated atypical bacteria. However, two crucial aspects merit consideration: firstly, the presence of atypical bacteria on nasopharyngeal swabs does not definitively establish them as the causative agents of pneumonia and the presence of typical bacteria should also be taken into account. Secondly, these studies were conducted in the early 1990s when the prevalence of macrolide-resistant pneumococci was significantly lower than today, explaining the favourable response to these antibiotics even in cases of pneumonia caused by typical bacteria [64, 65]. Indeed, macrolide-resistant pneumococci were almost lower than 10% before 1990 worldwide, but it started to increase, reaching more than 20–30% in late 1990 [66]. Over time, the improper and excessive use of macrolides has led to the emergence of resistance mechanisms in S. pneumoniae, the primary pathogens responsible for CAP [67]. In 2022, in Italy, over 25% of S. pneumoniae strains were resistant to macrolides [66].
The quality of the available evidence is low, due to the observational nature of most of the studies and the uncertainty of the diagnosis, and there are no studies found specifically addressing subjects unimmunized for H. influenzae type b.
Given S. pneumoniae’s resistance mechanism to penicillin, attributed to the production of penicillin-binding proteins rather than beta-lactamases [68], a robust recommendation supports using amoxicillin in subjects immunized for H. influenzae type b but not for S. pneumoniae. Conversely, H. influenzae’s resistance mechanism to penicillin stems from the production of beta-lactamases [69]. Consequently, antibiotics coupled with inhibitors of these enzymes become necessary, as evidenced by studies on other infectious diseases like acute otitis media. However, since there is a dearth of studies directly involving patients unimmunized for H. influenzae type b, thus lacking comprehensive data on the actual incidence of pneumonia infections from this pathogen in such subjects, the recommendation for the use of amoxicillin-clavulanate in these patients is weak, primarily extrapolated from other pathologies such as acute otitis media.
Question 4. What is the first-line antibiotic in the treatment of mild-moderate bacterial CAP in patients allergic to penicillin?
Recommendation 4
In patients with CAP and a suspected allergy to amoxicillin, who have not undergone allergological workup, the selection of alternative antibiotics (such as third-generation cephalosporins or macrolides) should be guided by meticulous risk stratification. (Very low-quality evidence. Expert opinion. Strong recommendation in favor of the intervention. Panel consensus 100%).
In patients with CAP suspected of having an allergy to amoxicillin and deemed to be at low risk of allergic reaction, a second or third-generation cephalosporin (such as cefuroxime or cefpodoxime proxetil) is recommended as an alternative therapy. The utilization of macrolides (like clarithromycin) or clindamycin should be reserved for patients at high risk of allergic reaction, with consideration given to levofloxacin for older children (Very low-quality evidence. Expert opinion. Weak recommendation in favor of the intervention. Panel consensus 100%).
Literature regarding penicillin allergy remains scarce (Table 1). β-lactam hypersensitivity reactions can be classified based on the time between antibiotic intake and symptom onset into immediate (symptoms within 1 h) and delayed (after 1 h). Immediate reactions are usually IgE-mediated and present as anaphylactic shock, angioedema, urticaria, and bronchospasm. Non-immediate reactions are typically non-IgE-mediated, with the most common clinical manifestation being a maculopapular rash, which is typically non-pruritic [70,71,72,73].
The risk of cross-reactivity between penicillin and cephalosporins remains a contentious issue [74]. Non-recent studies have reported a cross-reactivity frequency of 10% of cases. However, these studies were affected by evident biases, including the definition of allergy based solely on clinical history, the utilization of reference preparations of cephalosporins that were still “unrefined” and potentially contaminated by penicillin, and a lack of precise knowledge regarding the molecular structures of the compounds involved. Studies utilizing monoclonal antibodies have further revealed that for cephalosporins, the principal antigenic determinants are represented by the side chains. Therefore, in cases where the side chains differ from those of penicillin or amoxicillin, an increased risk of cross-reactions has not been demonstrated [75]. Supporting this, cefuroxime (a second-generation cephalosporin) has frequently been observed to be safe in patients with confirmed hypersensitivity to other beta-lactams, with cross-reactions occurring in only 6.3% of cases. More frequent cross-reactions have been noted with first-generation cephalosporins, which possess a beta-lactam ring similar to penicillins [76]. Given the low cross-reactivity observed in various studies, Drug Provocation Tests (DPT) are often conducted with cephalosporins featuring different side chains.
In non-IgE-mediated forms, T-lymphocytes play a role, and therefore, cross-reactivities are even rarer. In fact, 97.2% of individuals with hypersensitivity to aminopenicillins tolerate cephalosporins with different side chains, suggesting tolerability in the majority of cases [75]. Nonetheless, if penicillin allergy is confirmed, DPT with cephalosporins featuring different side chains may be considered.
In various allergy guidelines [70,71,72,73], there is no definitive consensus on the optimal method to ascertain if a patient with penicillin allergy can safely take a cephalosporin with different side chains directly or following DPT. However, the allergological assessment for a patient with penicillin allergy should ideally culminate in recommendations regarding the potential use of other beta-lactams, particularly cephalosporins. The determination of whether such use is feasible serves as guidance for family pediatricians in their future therapeutic prescribing practices.
Although macrolides therapy showed efficacy in treating CAP [31, 46, 64, 65], it’s important to note that most of these studies were conducted in the early 1990s when the prevalence of pneumococcal resistance to macrolides was significantly lower than it is today. One study in pediatric population showed that levofloxacin therapy was non-inferior to beta-lactam or macrolide therapy in terms of clinical cure rates, without difference in serious adverse events [61].
Considering the limited scientific evidence available, children with CAP and penicillin allergy, whether IgE or non-IgE mediated, may be treated with cephalosporins with different side chains (such as cefuroxime or cefpodoxime proxetil) only after undergoing an adequate allergological workup to assess their potential use. Alternatively, depending on the local prevalence of pneumococcal resistance to specific drug classes, clindamycin, clarithromycin, or, in older children, levofloxacin can be utilized [72].
Question 5. What should be the optimal dosage of amoxicillin in treating mild to moderate bacterial CAP?
Recommendation 5
To treat mild-moderate CAP, we recommend administering amoxicillin at a dosage of 80–90 mg/kg/day divided into three separate doses (with a maximum of 1 g three times a day). However, to enhance compliance with antibiotic therapy, particularly in cases of mild pneumonia with close clinical follow-up, the number of daily administrations can be reduced to two instead of three. (Moderate quality of evidence. Weak recommendation in favor of the intervention. Panel consensus 100%)
The available data regarding the optimal dosage of amoxicillin for treating CAP are very limited (Table 1). The study conducted by Bielicki et al. [38] suggests the non-inferiority of low-dose therapy compared to high-dose therapy, administering the antibiotic twice a day instead of three times. Despite the robustness of the trial, intrinsic limitations may introduce bias. As the authors noted, it remains unclear how many enrolled children had bacterial pneumonia versus viral pneumonia that might have resolved spontaneously without antibiotic therapy. Although efforts were made to limit bias by excluding children with acute bronchospasm lacking clear signs of pneumonia, the uncertainty persists. Additionally, the median age of the participants was notably low, under three years, a period when most CAP cases are viral. In such instances, antibiotics may be unnecessary, suggesting that the lack of difference in treatment failure between the low and high dose groups could reflect the low incidence of bacterial pneumonia in this age group rather than the efficacy of the low dose.
Moreover, this study was conducted in a setting with a low prevalence of S. pneumoniae resistant to penicillin and amoxicillin. Since pneumococcal resistance to penicillin is attributed to genetic mutations that alter penicillin-binding protein structure, resulting in decreased affinity for all beta-lactam antibiotics, increasing the dosage of penicillin to saturate the binding sites becomes necessary to counteract this resistance mechanism. The other studies included in this consensus primarily report the use of high-dose amoxicillin administered in three doses for the treatment of community-acquired pneumonia [39, 41, 45]. However, these studies do not compare different dosages in terms of therapeutic efficacy and the risk of adverse effects It is therefore plausible that in regions with a low prevalence of resistance, therapy with reduced doses of amoxicillin may be equally effective compared to higher doses. However, considering the specific resistance mechanism, the differences in antimicrobial resistance between Northern and Southern Europe, and awaiting further studies, the panel’s decision in our region is to continue recommending therapy with high-dose amoxicillin (90 mg/kg/day), administered in three daily doses. Nevertheless, the strength of recommendation is weak, as lower dose might be efficacy as well. With close clinical monitoring, there is potential to reduce the number of daily administrations to two instead of three to enhance adherence to antibiotic therapy in cases of mild pneumonia.
Question 6. What should be the optimal length of therapy with amoxicillin for treating mild -moderate bacterial CAP?
Recommendation 6
For the management of mild-moderate CAP, a 5-day course of antibiotic therapy with amoxicillin is recommended. Close clinical monitoring and reassessment are advised approximately 72 h after initiating antibiotic therapy to evaluate symptom resolution. If necessary, treatment may be extended for up to 7 days. (Moderate quality of evidence. Strong recommendation in favor of the intervention. Panel consensus 100%)
Based on the analysis of scientific evidence derived from four different randomized control trials [38,39,40,41], four systematic reviews [32,33,34,35], and three observational studies [45, 47, 53], it has been established that short-term therapy is not inferior to long-term therapy in treating mild-moderate CAP. Long-term therapy, extending to 10 days or beyond, does not appear to offer any advantage in terms of reducing therapeutic failures compared to shorter durations of therapy. Consequently, it is not recommended to prescribe prolonged antibiotic therapy, unless complete clinical resolution is not achieved during follow-up assessments.
The duration of short therapy varies across studies, ranging from 3 to 7 days. However, the sole trial that compared a 3-day therapy regimen to a 10-day regimen was suspended due to an increase in therapeutic failures [41], whereas no difference was observed comparing a 3-day therapy regimen to a 7-day therapy regimen [38].
Given the challenge of definitively determining how many of the enrolled children in various studies had bacterial pneumonia versus viral pneumonia, which could have resolved spontaneously without antibiotic therapy, it is recommended to maintain close follow-up approximately 72 h after initiating antibiotic therapy, by phone call in case of improvement and reliable family or re-evaluation in person in case of persistence of symptoms or unreliable family. This allows for assessment of the need to extend antibiotic therapy if symptoms do not completely resolve.
Therefore, until further studies with more robust scientific evidence comparing 3 and 5 days of therapy are available, the panel recommends prescribing antibiotic therapy for 5 days to children with uncomplicated CAP. Close clinical follow-up and reevaluation should occur approximately 72 h after initiating antibiotic therapy to assess symptom resolution or the necessity for continuation up to 7 days if needed.
Question 7. What is the most appropriate antibiotic therapy in a child with CAP experiencing clinical deterioration after 48 h of first-line therapy with amoxicillin?
Recommendation 7
In children experiencing clinical deterioration after 48 h of first-line therapy, hospitalization and treatment with broad-spectrum antibiotics are recommended.
(Evidence quality very low. Weak recommendation in favor of the intervention. Panel consensus 100%)
No articles were identified in the literature regarding second-line therapy to be used in the community in the event of therapeutic failure and clinical deterioration. Several factors hindered the inclusion of additional studies comparing broad-spectrum and narrow-spectrum therapy in children hospitalized for CAP. Firstly, many of these studies also encompassed children with complicated pneumonia, a population not addressed in this consensus. Additionally, these studies often failed to specify whether hospitalization and the initiation of broad-spectrum parenteral antibiotic therapy were due to therapeutic failure of prior oral therapy or solely based on the child’s clinical condition. Consequently, it was challenging to understand whether children who had not improved with oral therapy were treated with narrow-spectrum or broad-spectrum therapy and what clinical improvement they had. Only one study was found regarding treatment failure of oral antibiotics, but children were hospitalized and treated with narrow-spectrum antibiotics (penicillin, ampicillin, amoxicillin) or broad-spectrum antibiotics (ceftriaxone, cefuroxime) with a statistically significant reduction in the duration of fever and length of stay in hospital in the broad-spectrum group [55].
Therefore, until further studies with more robust scientific evidence become available to recommend oral antibiotic therapy in cases of therapeutic failure associated with clinical deterioration, it is advised to hospitalize the child and initiate treatment with broad-spectrum antibiotics intravenously.
Add Comment