Definitions and Outcomes of Nutritional Interventions in Children with Respiratory Infections: The Approach of the COMMENT Initiative Free

Acute respiratory infections (ARI), including both upper and lower respiratory infections (URTI and LRTI), are the most common illnesses worldwide during infancy [1]. The high rates of respiratory infections are associated with high social and family costs [2].

ARI are a major cause of missed work-days by parents and are responsible for a massive use of drugs and investigations.

Risk factors for ARI, in addition to host-related conditions, include environmental conditions such as seasonality (i.e. wintertime with higher rates of influenza, respiratory syncytial virus and other viruses) and selected settings such as daycare centers, schools and hospitals. Children attending daycare centers are at a 2- to 3-times greater risk for developing ARI than children at home [3]. Host-related conditions such as immune-compromising conditions, underlying chronic diseases and atopy are associated with an increased incidence and severity of ARI [4].

Scattered data suggest that selected nutritional interventions may reduce the risk of developing ARI. Probiotics, prebiotics, vitamins and micronutrients such as zinc have been used to reduce the incidence of gastrointestinal and respiratory infections in pediatric populations. Results are sometimes conflicting and often difficult to generalize. This may depend on factors such as population features (e.g. age, location, health status, risk factors and other therapies) and type of intervention (e.g. type of supplementation, formulation, dosage and duration of consumption), but may be even more strongly related to the difference in definitions and outcome measures that have been used.

The recently formed Consensus Group on Outcome Measures Made in Paediatric Enteral Nutrition Clinical Trials (COMMENT) appointed specific working groups to identify and define criteria for assessing key outcomes in these trials. The overall aims of the COMMENT initiative are reported in detail in a previous paper by Koletzko et al. [5].

Inspired by the COMMENT initiative, a panel of experts including members of the European Society of Pediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) and the European Society of Paediatric Infectious Diseases (ESPID) decided to conduct a systematic review of the available data, to define the most appropriate criteria to apply in trials that study the effects of nutritional interventions for the prevention or treatment of ARI in children [5].

This study is aimed at reviewing the data from the trials, looking at the effects of selected nutritional interventions on ARI in infancy and early childhood, with specific attention paid to the definitions, clinical end points and markers used.

We searched for clinical trials studying the effects of different nutritional interventions on ARI in children and infants.

A research on MEDLINE (up to August 2012) and on the Cochrane Central Register of Controlled Trials (CENTRAL 2012) databases was performed, with no limit on article type or time, in order to obtain a full body of evidence to be selected afterwards.

Search terms included extensive controlled vocabulary and keyword searches for ‘respiratory tract infections' [Medical Subject Headings (MeSH) and text words (tw)], ‘influenza' (tw), ‘otitis' (MeSH and tw), ‘pneumonia' (MeSH and tw), ‘infants' (MeSH and tw), ‘child*' (MeSH and tw), ‘nutrition', ‘infant formula' (MeSH), ‘yoghurt' (MeSH and tw), ‘probiotics' (MeSH), ‘prebiotics' (MeSH), ‘synbiotics' (MeSH), ‘micronutrients' (MeSH), ‘calcium' (MeSH and tw), ‘vitamin' (tw), ‘lipid' (MeSH), ‘fat*' and ‘fats', ‘carbohydrates' (MeSH) and ‘protein' (MeSH).

To identify additional articles, references cited in the included trials were checked.

We decided to include studies conducted on children up to the age of 6 years (preschool age), published in English, that include URTI or LRTI as primary or secondary outcomes, or biomarkers related to respiratory tract inflammation or infection episodes. Although the COMMENT project included infants and young children aged <3 years [5], we decided to extend the age range up to 6 years with the ultimate aim of including preschool children who more frequently develop respiratory infections and are the usual target population for such trials on respiratory infections. Any type of nutritional intervention was considered. However, administration of ‘functional foods' as such, including prebiotics, probiotics or synbiotics not included in formulas or other foods, was not included.

Trials on populations with a wide age range including both target population (infant and preschool children) and older children were included, even if age-specific subanalysis was not reported. Trials that included only children ≥6 years of age were excluded.

Articles were selected based on the title and abstract, and the full text of all selected articles was retrieved and independently assessed for inclusion according to prespecified criteria by two independent reviewers (E.B. and A.L.V.).

A standardized table of evidence was prepared including author, year of publication and journal, type of trial, target population, interventions, controls, primary outcomes and definitions of respiratory illness, clinical outcomes and the assessment of biomarkers to measure the effects of nutritional interventions on respiratory outcomes. In articles including respiratory and nonrespiratory clinical outcomes, only data related to respiratory clinical features and related definitions were reported in the table of evidence.

Descriptive statistics were used to describe the characteristics of trials and compare definitions used by researchers, interventions and outcomes.

The electronic and bibliographic research identified 107 suitable references. After abstract screening and application of study inclusion criteria, 50 papers were included, 46 focusing on the prevention of respiratory episodes and 4 on treatment [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55]. Figure 1 shows the flow-search diagram.

Trials included in our analysis were published between 1991 and 2012 (21% between 1991 and 1999 and 79% between 2000 and 2012). More than 90% were specifically conducted to evaluate the effects of nutritional interventions in the prevention or treatment of respiratory tract infections in children, with 10% including respiratory end points as secondary outcomes.

The age of enrolled children ranged from 15 days to 15 years.

Fifteen studies (30%) included only infants, 28 (56%) included children up to 5 years. Others included children with wider age ranges including infants and older children (5 with an age range of 1 month to 15 years) or ages were not specified [2].

The majority (approx. 70%) of trials were conducted on healthy subjects. However, several studies evaluated the efficacy of specific nutrients in children with clinical conditions such as malnutrition, HIV infection or atopy as well as children who had been hospitalized, which potentially increases susceptibility to infectious diseases. Other important factors which could modify the risk of infectious disease were the socioeconomic status and the social setting. Both may influence nutrient availability, health assistance, prevention measures and case exposure. There was a community setting in 68% (34/50) of the trials; the rest investigated the effects of nutritional intervention in high-risk settings such as hospitals, daycare centers and schools. We included trials conducted in different locations, i.e. Africa: 9 (18%), North America: 5 (8%), South America: 7 (14%), Asia: 15 (31%) and Europe: 14 (29%). The sample size was very broad, ranging from 40 to 15,419 and the duration of the intervention ranged from days to several months (≤6 months in 68% of the trials).

The nutritional interventions applied in the analyzed studies were broad and included infant formulas and yoghurt (enriched with various prebiotics or probiotics and/or micronutrients) and supplementation with vitamins and micronutrients (table 1).

The definitions of respiratory infections were highly heterogeneous.

A specific segmental definition of URTI or LRTI was reported in 45/50 (90%) trials. In 5, no definition of respiratory infections was reported. In 15 of the 50 analyzed trials, the definitions of URTI, LRTI and acute otitis media (AOM) were based on a specific diagnosis made by a pediatrician (rhinitis, laryngitis, tracheitis, pharyngitis, sinusitis, otitis or common cold/influenza). In 30 trials, the definition of URTI or LRTI was based on clinical symptoms reported by families or field-workers (runny nose, cough or sore throat). LRTI included pneumonia, bronchitis, wheezing and bronchiolitis. A specific diagnosis based on cough, abnormal respiratory rate according to age, crepitations to chest auscultation and drawing in of the chest was made in most cases, and this was supported by radiographic findings in some.

In some trials, other systemic symptoms and signs such as fever, headache, restlessness, aphonia, shortness of breath and acute ear pain were added to the respiratory features to further support the diagnosis. Fever was reported in 22 out of 50 trials reported as a feature associated with URTI or LRTI, although a specific definition of fever was provided in 8 papers (36%) only, and the cut-off temperature values varied with a rectal temperature of >38°C being the most common definition. The temperatures were usually reported by parents or field-workers, and the duration of each episode was recorded in days (rather than in hours).

The most common, specific definitions are reported in table 2.

The duration of episodes was considered as a major outcome of respiratory infection episode in 8 of the trials (3 were on treatment and 5 were on prevention). In 4 trials (50%), an episode started 24 h after the onset of symptoms, i.e. very brief or transitory episodes were excluded and in the other papers, at least 2 consecutive days of symptoms were needed to define a single episode. Recurrence was another important outcome in 19 trials, and a new episode was defined as coming after a symptom-free interval of 2, 7 or 14 days in 1, 2 or 4 trials, respectively.

Incidence, prevalence and duration of specific respiratory symptoms (e.g. cough with or without fever) were the main outcomes in trials on prevention. Duration of symptoms or hospitalization and symptom-free periods were the main outcomes in the 4 trials on treatment (table 3).

Some trials included quantitations of antibiotic prescriptions, absence from daycare or school, or even medical visits as surrogate end points, if no specific definition was provided (table 4).

Nonspecific markers related to respiratory tract infections or inflammation were also a topic in some of the studies. Some reported nonspecific laboratory markers of inflammation (e.g. C-reactive protein, white blood cell count, immunoglobulins, interleukins and retinol-binding protein) as secondary outcomes.

This study aimed to systematically review the definitions and outcome measures used in trials assessing the efficacy of nutrition in the prevention and treatment of respiratory tract infections in infancy and early childhood. The results showed a major heterogeneity, with the use of a wide array of definitions and clinical end points.

Such variability in defining disease episodes and end points could be partially related to the important role played by family members and/or field-workers in reporting conditions, particularly in the large field-trials on prevention. In this specific setting, definitions need to be broad and easy to measure, with the risk of losing some sensitivity but mainly specificity.

Different definitions of URTI and LRTI were used based on the presence and association of specific symptoms; however, the duration of these was rarely reported. Diagnosis of URTI was often based on a single symptom or on the presence of a cohort of different respiratory symptoms (runny nose, nasal congestion, laryngitis, pharyngitis or tracheitis). The association and the duration of the symptoms varied between studies. In most cases, the presence of symptoms was recorded by a member of the family of the sick child (usually the mother), but in some, the definition was based on the diagnosis made by the physician and a specific (segmental) definition of URTI, rather than specific symptoms, was reported.

Criteria for the diagnosis of AOM were also poorly reported and heterogeneous throughout the studies. Tympanic membrane bulging is a key sign to differentiate AOM from otitis media with effusion [56], and the assessment of membranes through a pneumatic otoscopy is a prerequisite for the diagnosis. However, this requires a medical staff well-trained in pneumatic otoscopy [57], and may represent a barrier to including otitis among the outcome measures for large-scale studies.

In the case of LRTI, the definition was somehow more consistent in the few articles that included it, also because the diagnosis was based on physician consultation and specific symptoms of LRTI (such as the increase of respiratory rate, crepitations to chest auscultation or drawing in of the chest) or else on radiographic findings.

Fever was frequently taken as an outcome measure, being easy to measure and monitor as well as providing information on the duration and, to some extent, on the severity of the disease. However, the reliability of body temperature measurements may vary widely according to instrument, environment and site of measurement. In addition, the definitions of fever were scattered with cut-off values ranging from 37 to 38.5°C.

Overall, most studies focused on prevention, the heterogeneity for primary outcomes was limited and the incidence of new URTI or LRTI episodes was the outcome measure most frequently used, despite the definitions not always being sufficiently detailed. In the majority of trials, important information such as time limits of intervention and observation was not included. Other potentially relevant details were often missing. Reference to the season, which can be a key risk factor for developing respiratory infections and usually included in the definition of community-acquired respiratory infections (e.g. winter season to define influenza or other respiratory viruses) was rarely found. Moreover, the wide age ranges of the study populations and the different study settings created major potential biases which affected the accurate assessment of respiratory infection outcomes. Some relevant markers specifically related to clinical outcomes, e.g. vaccination status with regard to influenza or Streptococcus pneumoniae, were not provided in the majority of cases that had been conducted after licensure of these vaccines.

Considering the current scenario and the relevant heterogeneity reported, a straightforward definition of outcome measures seems to be needed in order to ensure a more reliable and consistent reporting of data.

We hypothesized differentiating outcomes into ‘direct' and ‘indirect' outcomes. The ‘direct outcomes' would be aimed at assessing the efficacy of a selected intervention on respiratory diseases. These outcomes, including the incidence or the severity of selected infections (e.g. otitis, URTI and pneumonia), should be measured by well-trained personnel (physicians) who make a specific diagnosis.

On the other hand, the ‘indirect outcomes', e.g. the number of chest X-rays performed, medical visits and interventions or hospitalizations as well as the loss of workdays, may provide a reliable estimate of the burden of respiratory diseases on healthcare. These simple, easy-to-measure end points may be monitored by field-workers or even family members (if trained), thereby bypassing the need for difficult or complex diagnostic criteria or validated scores.

In conclusion, heterogeneity was revealed in the current literature on respiratory infections and problems arose when we compared the results of studies using heterogeneous definitions and end points. These factors highlight the need for ‘guidelines' that agree on standard definitions and outcome measures in order to facilitate trials in the future that will be effective in assessing respiratory outcomes in childhood.