Data sources and search strategy

Relevant peer-reviewed journal databases (PubMed, Medline, Web of Science, ScienceDirect (Elsevier), OvidSP and EBSCO host) were searched using the following MeSH (Medical Subject Headings) terms and keywords: “climate change”, “humidity”, “wet”, “dew point”, “child”, “infant”, “adolescent” and “pediatric”. The search terms and keywords were adapted to suit the requirements of each database to facilitate a comprehensive search. After selection of the full text, the references of the identified articles were reviewed and the necessary hand search was performed to recover potentially relevant studies not included in the major databases.

The search sought to identify journal articles published in English up to March 2013. In accordance with the Convention on the Rights of the Child proclaimed by the United Nations, we defined children as the group of people 0–18 years of age [17].

Inclusion criteria

The following eligibility criteria were used in this study:

• Types of studies: the present review only included peer-reviewed original articles. Thus, reviews, books, meta-analyses and conference abstracts were excluded.
• Target population: only the studies involving children between 0–18 years of age were considered.
• Research factors: only studies in which humidity was one of the exposure indicators of interest were considered.
• Outcome measures: the reported results of the included studies should have included at least one quantitative effect estimate associated with humidity. An outcome measure was reported as either: percent change, regression coefficient, correlation coefficient, relative risk or odds ratio. We also required P-values or 95% confidence intervals (CIs) to have been reported.

Study selection and data extraction

Three of the authors (Gao, Sun and Lu) independently assessed the potential relevance of all publications, identified during the database search, based on the information provided in the titles and abstracts. When reviewers disagreed, a discussion was held to obtain consensus. After screening by titles and abstracts, the full texts of the remaining articles were then obtained and critically reviewed by the first author.

We extracted data, from the final selected references, using a pre-designed data extraction form. The quantitative estimates of each study that met the eligibility criteria were carefully retrieved and the following key characteristics were captured: study region, target population, study design, humidity exposure variable, health outcomes and adjustment of confounders. All data extraction was performed by the first author (Gao) and cross-checked by another researcher (Sun). Any errors identified were corrected through discussion among all authors.

Quality assessment

To date, the use of quality assessment tools to appraise observational studies included in systematic reviews is less well established than in reviews of randomized controlled trials, and there is no existing generally accepted quality assessment tool for observational and cross-sectional studies [18], [19]. Therefore, the STROBE method (Strengthening the Reporting of Observational Studies in Epidemiology), a guideline for reporting observational studies, was selected and adapted for our purpose to describe the reporting quality of the final identified articles [20], [21].

As in a previous systematic review, an adapted version of the STROBE checklist (Table 1) was used to create a binomial scoring system (e.g. criterion met  = 1, criterion not met  = 0). Final score percentages were based on the number of reporting quality criteria met divided by the number of criteria that were relevant during the quality assessment [22]. During this process, studies that controlled for potential confounding factors and included an outcome measure, such as relative risk, odds ratio or regression coefficient, were given higher quality ratings. Meanwhile, studies were given lower quality ratings if they did not include adjustment of confounders or only included outcome measures such as an estimated slope coefficient from simple linear regression or a correlation coefficient for which only the P-value was given. As no accepted gold standard exists for judging the quality of observational studies, in the present review we did not exclude any identified studies based on the quality assessment, which used an adapted version of the STROBE checklist.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Adapted STROBE Statement—checklist of items that should be included in reports of observational studies (including additions/adaptations for accommodating meteorological data).

https://doi.org/10.1371/journal.pone.0112508.t001

Data synthesis

Due to heterogeneity in the target populations, outcome measures and design of studies, we did not conduct an overall summary estimate (e.g. meta-analysis). Instead, we summarized the evidence in a descriptive synthesis.

Gastrointestinal and other infectious diseases

Studies found that ambient humidity plays an important role in the occurrence and prevalence of pediatric infectious diseases. A Brazilian study found that among the children hospitalized with community-acquired pneumonia, overall viral and Chlamydia trachomatis infections correlated with monthly mean relative humidity [23]. In Taiwan, a study identified a positive association between daily mean relative humidity and the incidence of enterovirus 71 (EV71) infections among children under 15 years of age (IRR = 1.08, 95% CI: 1.05–1.10), with seasonal and annual trends being controlled [24].

In Japan, one study indicated that for each 1% increase in the weekly mean relative humidity, the number of mumps cases increased by 1.4% (95% CI: 0.5–2.4%) among children [25]. In another Japanese study, the weekly number of hand, foot, and mouth disease cases increased by 4.7% (95% CI: 2.3–7.1%), in children between 0–4 years of age, for each 1% increase in relative humidity [26]. An additional study investigated the relationship between weather variability and infectious gastroenteritis (IG) in children. However, the results indicated that for every 1% decrease in weekly mean relative humidity, the number of IG cases increased by 3.9% (95% CI: 2.8–5.0%) [27].

Existing studies suggest that ambient humidity has a significant impact on malaria risk and diarrhoeal diseases among children. A prior study found that the clinical malaria risk increased non-linearly with the increase in mean relative humidity of the previous month [28]. A Mozambican study reported that humidity was a significant predictor of malaria incidence, with every additional 1% increase in relative humidity level leading to a 5% increase in risk of malaria incidence [29]. In Taiwan, a study examining the relationship between climate variations and diarrhea-associated morbidity found that there was a significant negative correlation between relative humidity and the morbidity of diarrhea in 0- to 14-year-old children (β = −0.033, P = 0.004) [30]. Another study performed in Peru indicated that an increase of 1% in mean relative humidity was associated with a 3% decrease in the number of diarrhoeal admissions [31]. The results of an Australian study showed that higher humidity in the previous week was associated with a decrease in rotavirus diarrhoeal admissions among children hospitalized for rotavirus diarrhoea after controlling for season [32].

Respiratory system diseases

Many pediatric respiratory system diseases are sensitive to relative humidity. An Argentinean study reported that respiratory syncytial virus (RSV) and influenza virus A had a positive correlation with mean monthly relative humidity among children hospitalized due to acute lower respiratory viral infections [33]. Another study found that low absolute humidity presented a protective effect on the admission of infants who had lower respiratory tract infection (LRTI) due to RSV [34]. In a study conducted in Jordan, a significant positive correlation between the prevalence of RSV infection and relative humidity was identified among hospitalized children [35]. A similar study in Indonesia indicated that for children hospitalized with clinical pneumonia, a 1% rise in mean relative humidity was associated with a 6% increase in RSV cases [36]. An additional study of children hospitalized with LRTI in Turkey reported that the number of RSV infection cases positively correlated with relative humidity, similar to results reported by Noyola and Mandeville for Mexico [37], [38]. Another retrospective study was performed in Malaysia to investigate the viral etiology of children who were admitted with viral respiratory tract infections. However, the results showed that monthly RSV cases were inversely correlated with relative humidity [39].

In Germany, a study found that human rhinovirus positively correlated with relative humidity among children who were hospitalized with acute respiratory tract infections (ARIs) [40]. Another study in New Zealand showed that the number of days per month with a mean 9 AM humidity >75% positively correlated with the incidence of invasive pneumococcal disease in children [41]. The results of a study in Cameroon showed that high relative humidity was directly associated with an increase in the frequency of hospitalization from ARIs [42]. However, in one Singapore study, of 42155 children presenting in the hospital with one of five common paediatric diseases, both upper respiratory tract infections and LRTI were negatively correlated with relative humidity [43].

Dermatopathy and pediatric allergic diseases

Childhood asthma and other pediatric allergic diseases are important adverse consequences of humidity. In Greece, the results of a study showed that relative humidity was an implicated meteorological variable for younger asthmatic children [44]. A study performed in fourteen cities in Turkey found that asthma, wheezing and allergic rhinitis were associated with higher mean yearly outdoor humidity among school children [45]. Another study reported an interesting finding that relative humidity was significantly associated with the incidence of acute urticaria in childhood. However, the humidity accounted for different effects in the two study cities. Specifically, decreasing incidence was found for humidity increases in Norwich, Britain, while the inverse was observed in Heraklion, Greece [46].

In one Turkish study of 21 bronchial asthma children, results indicated that increasing relative humidity leads to an increase in total daily complaints of asthma patients [47]. A study reported that winter humidity was positively associated with the prevalence of asthma in Taiwanese middle-school students [48]. A positive relationship between mean monthly relative humidity and childhood asthma admissions (CAAs) was reported in Greece. For this study, a 10% increase in humidity was related to a 31% increase in the probability of having a CAA [49]. In addition, an analysis, carried out exclusively within the coastal region of Spain, showed that relative humidity constituted a significant risk factor for wheezing in both the 6–7 and 13–14 age groups [50]. A retrospective study performed in America indicated that a 10% intraday increase in humidity one or two days before admission was associated with approximately 1 additional emergency department visit for asthma [51].

On the contrary, a study in Taiwan indicated that relative humidity was negatively correlated with asthma admission rates of pre-school and school-age children [52]. In Western Europe, a positive relationship between the lowest monthly mean relative humidity and prevalence rates of asthma symptoms was found among children [53]. Another similar study conducted among school children in Taiwan indicated that the lowest monthly mean relative humidity is positively related to eczema [54]. An additional study conducted in Japan reported that the number of emergency visits for childhood asthma per night increased significantly when climate conditions showed a rapid decrease from higher humidity [55]. A study in Spain also showed that relative humidity was negatively associated with asthma in the 13–14 age group [56].

Others

Dew point, another measure of ambient humidity, is the temperature at which water vapor in the air becomes saturated and water droplets begin to form. To date, relative humidity and dew point are two widely used indicators for the amount of moisture in air. Accordingly, some other studies also investigated the association between ambient humidity and child health by choosing dew point as the indicator. A negative association between the peak week of child-related influenza visits and dew point was reported by Charland and colleagues, although the relationship was not significant [57]. In another study, Wang et al. found that a higher dew point was associated with a lower risk of hospital admission among children [58].

In addition, ambient humidity may also play a role in other diseases among children. For instance, the results of an American study showed a positive association between relative humidity and the likelihood of headache among children [59].

Impact of ambient humidity on child health

Because of their physiologic, metabolic and cognitive immaturity, and greater sensitivity to certain exposures, children are often the most vulnerable to adverse health effects from environmental hazards [14], [16]. In addition, more expected future years of life will expose them to newly developing or worsening environmental hazards [7]. In general, child vulnerability to environmental hazards (e.g. humidity) can be potentially explained by their unique modality of physiological, metabolic, behavioral, and self-care ability [60], [61]. Therefore, with global climate change, children are likely to suffer disproportionately from weather factors [60], [62]. It is thought that if children are adversely affected by meteorological factors (e.g. humidity), their long-term health may also be adversely impacted [63].

Findings from the present review suggest that the occurrence and prevalence of childhood climate-sensitive infectious diseases are significantly associated with ambient humidity. It is predicted that climate change will lead to more frequent and abundant rainfall, increasing temperature and higher concentrations of water vapor, which will cause favorable conditions of ambient humidity and concomitantly favor increases in the occurrence and development of infectious diseases [3], [4], [10]. For example, ambient humidity together with temperature can influence the population dynamics of vectors, persistence and survival of pathogens, seasonal activity of microorganisms, and duration of the life cycle of parasites, and therefore impact the occurrence and prevalence of various infectious diseases [10], [63], [64].

Ambient humidity can affect tick abundance and seasonal tick activity. It has been reported that the host seeking activity of I. ricinus depends on the relative humidity, and below a relative humidity threshold of 86%, the tick dehydrates and cannot continuously seek a host [65], [66]. In Sweden, a study found that the number of days with levels of relative humidity above 86% correlates positively with the incidence rate of erythema migrans [67].

For gastrointestinal infectious agents such as rotaviruses and EV71, changes in ambient humidity have been reported to facilitate viral resistance, thereby enhancing the virulence and increasing the efficiency of transmission through contaminated air, fomites and environmental surfaces [68], [69]. A laboratory-based study suggested that the stability of enteroviruses in external environmental conditions is dependent on temperature and humidity, and another study reported that raising the relative humidity to 80% results in a rapid loss of infectivity, suggesting that humidity might be an important environmental determinant of rotavirus infectivity [68], [70]. This perhaps can explain why the relative humidity has a linear inverse relationship with the number of cases of rotavirus diarrhea previously reported in a study of 3115 rotavirus diarrhoea cases in 0- to 2-year-old Bangladesh children [71]. However, a study performed in China suggested that the number of hand, foot, and mouth disease infections is positively correlated with ambient humidity [72].

Malaria is a climate-sensitive vector-borne illness. Meteorological factors (e.g. temperature, humidity) have considerable impact on Anopheles vector abundance and the extrinsic cycles that the parasites perform inside mosquitoes. Thus, meteorological factors may affect malaria incidence and constitute driving forces for malaria epidemics [73]. A study in China suggested that moderate temperature and relative humidity increase the longevity of the adult mosquito so that it can transmit infection for a longer period of time [74]. In India, a study found a higher positive correlation between monthly malaria parasite incidence and ambient humidity [75].

Many studies demonstrate the effects of relative humidity on pediatric respiratory tract infections. In tropical and subtropical areas (e.g. Singapore, Malaysia), it has been suggested that high humidity and temperature prolong the activity and stability of microorganisms in large-particle aerosols, and permit year-round transmission of the microorganisms [5], [76]. A study reported that relative humidity is positively associated with RSV activity, with more RSV infection occurring when humidity increases [77]. However, a converse conclusion was found in an Argentinean study, which concluded that humidity provides a protective effect against respiratory infections [78].

The possible effects of ambient humidity on childhood asthma and other allergic diseases are of particular concern due to this population's greater susceptibility. Meteorological factors (humidity, temperature) can influence both biological and chemical components of air and might induce negative effects on respiratory allergic diseases. With climate change, ambient humidity is likely to impact asthma, hay fever and other allergic diseases via its influence on the concentration and distribution of pollens, mould spores, house dust mites, and other aeroallergens [79], [80]. Moreover, changes in the outdoor humidity, resulting from climate change, is important not only because of its own effects, but also because of its influence on indoor humidity, which results in infestation of homes with house dust mites, growth of fungal spores and home dampness, and is associated with the incidence of allergic diseases [12], [79], [81]. A study conducted among school children in Spain indicates that humidity significantly influences the occurrence of atopic eczema [82]. Youssefagha and colleagues indicate that upper-air relative humidity ≥50% and higher dew point (>−1.9) are significantly associated with exacerbation of asthma among children [83]. Another study performed in Japan also suggests that asthmatic children frequently visit the emergency department during early mornings with high absolute humidity [84].

However, in this review, some studies also report a negative association between ambient humidity and pediatric allergic diseases [52], [56]. The reasons for the inconsistencies among different studies are possibly due to regional differences and local characteristics, such as geographical environment, type of climate, vegetation condition, air quality, socio-economic status, and the acclimatization and adaptation of the local populations [63], [79], [80]. Additionally, study design, subgroups of children and method of analysis may also contribute to the controversies.

As mentioned above, different studies may present inconsistencies in the direction and magnitude of the effects of ambient humidity, with other studies even concluding that the occurrence of relevant climate-sensitive diseases appears to be unrelated to humidity [36], [39], [85], [86]. This discrepancy can be explained by the following possible reasons: firstly, relative humidity can affect microorganisms in different ways. Some microorganisms tend to be more invasive in increased humidity, while others develop better in lower humidity [10]. Similar effects of ambient humidity on the concentration and distribution of aeroallergens have also been reported [87]. Secondly, there is a geographical, meteorological and temporal heterogeneity in the different studies. The occurrence of different effects suggests that geographical sites, regional variability and acclimatization of the local population may play a role in determining the exposure–response relationship. Thirdly, among various studies, whether or not one takes into account potential confounding factors and what kind of confounder is adjusted for existing differences, can make a difference in the conclusions. Moreover, even mutual confounding factors among different studies are likely to vary between regions, which implies that regional differences and local characteristics may modify the effects of weather factors [25], [27], [71]. Lastly, the date of the study conducted, the limited study duration and participants, and the different age groups of the study population may also play a role in the discrepancy. Future research taking these factors into account will be appreciated.

Strengths and limitations

This systematic review begins to address the current knowledge gap in the relationship between ambient humidity and child health. This is the first attempt to synthesize, using a systematic approach, the existing evidence of humidity-related health outcomes, across different countries, among children between 0–18 years of age. According to Table 2, almost all of the final included journal articles (34/37) were published after the year 2000, and were of relatively high reporting quality (the majority of records having a reporting quality above 70%). The findings in Table 2 suggest that ambient humidity has been playing a significant role in weather-related health outcomes. Unfortunately, children are usually the most susceptible subpopulation. The present review is crucial because this literature provides details of the current state of studies that examine the effects of humidity on child health and proposes directions for the future research.

Despite targeting the reporting quality of observational studies, many items of the STROBE checklist were no doubt selected due to evidence of association with susceptibility to bias. Thus experts have suggested that the STROBE statement provides a suitable starting point for development of a quality assessment tool [20], [88]. Just as the MOOSE (Meta-analysis Of Observational Studies in Epidemiology) guidelines are being used to assess quality, although they were actually developed as reporting guidelines, in this review, we tried to describe the reporting quality of the final identified articles by using an adapted version of the STROBE checklist [18], [89]. After the quality assessments, studies that applied credible and convincing statistical methods and outcome measures often obtained higher quality scores at the same time (Table 2), indicating that the adapted STROBE checklist can describe the reporting quality of the final identified articles well. This novel quality assessment tool is innovative and potentially ground-breaking in terms of its contribution to relevant environmental and meteorological literature reviews [22].

However, regarding the results of this systematic review, some limitations should be acknowledged. First, this review is limited by its emphasis on papers published in English in peer-reviewed journal databases, so we may have missed potentially useful studies. Second, despite the highly sensitive and comprehensive nature of the search strategy employed, it is unlikely that all existing studies have been identified, and publication bias cannot be entirely excluded. Third, the adapted version of the STROBE checklist was initially implemented to describe the reporting quality of weather-related studies, so the validity and reliability of the checklist for our purposes have not been completely tested and formally evaluated. However, in this review, the tool was used only to describe the reporting quality, and the quality scores were not included as part of the relevance criteria. Fourth, it has been reported that temperature and relative humidity are not independent of one another, so the effects of relative humidity may be confounded by temperature. However, in our review, the majority of studies adjusted for potential confounding factors (e.g. temperature) when analyzing the impact of humidity on child health, and studies that conducted no adjustments were given lower quality ratings. Additionally, relative humidity is a popular indicator that has been widely used in current research and meteorological monitoring. Therefore, it is helpful and of practical significance to choose relative humidity rather than absolute humidity or water vapor pressure as the exposure variable. Finally, the data extraction and quality assessment were not performed independently by more than one reviewer, which may have led to subjective errors and introduced bias. However, in the present review, due to the explicit inclusion criteria, specially pre-designed data extraction form, definite exposure variable and health outcome, and the adapted quality assessment tool, the authors were able to ensure the integrity of the review, and to be unbiased during the conduction of assessment.