Association of meteorological parameters with intussusception in children aged under 2 years: results from a multisite bidirectional surveillance over 7 years in India

Objectives The study aimed to document the association between intussusception in Indian children and meteorological parameters and examine regional variations. Design A bidirectional (retrospective and prospective) surveillance between July 2010 and September 2017. Setting At 20 hospitals in India, retrospective case record review during July 2010 and March 2016 and prospective surveillance during April 2016 and September 2017 were performed. Participants 2161 children aged 2–24 months with first intussusception episode were included. Interventions The monthly mean meteorological parameters (temperature, sunshine, rainfall, humidity and wind speed) for the study sites were collected. Methods The association between monthly intussusception cases and meteorological parameters was examined at pooled, regional and site levels using Pearson (r) and Spearman’s rank-order (ρ) correlation, factorial analysis of variance, and Poisson regression or negative binomial regression analyses. Results The intussusception cases were highest in summer and lowest in autumn seasons. Pearson correlation analysis showed that temperature (r=0.056; p<0.05), wind speed (r=0.134; p<0.01) and humidity (r=0.075; p<0.01) were associated with monthly intussusception cases. Spearman’s rank-order correlation analysis found that temperature (ρ=0.049; p<0.05), wind speed (ρ=0.096; p<0.01) and sunshine (ρ=0.051; p<0.05) were associated with monthly intussusception cases. Poisson regression analysis resulted that monthly intussusception case was associated with rising temperature (North region, p<0.01 and East region, p<0.05), sunshine (North region, p<0.01), humidity (East region, p<0.01) and wind speed (East region, p<0.01). Factorial analysis of variance revealed a significant seasonal difference in intussusception cases for pooled level (p<0.05), 2–6 months age group (p<0.05) and North region (p<0.01). Significant differences in intussusception cases between summer and autumn seasons were observed for pooled (p<0.01), children aged 2–6 months (p<0.05) and 7–12 months (p<0.05). Conclusions Significant correlations between intussusception cases and temperature, humidity, and wind speed were observed at pooled and regional level in India. A peak in summer months was noted, which may be used for prediction, early detection and referral for appropriate management of intussusception.


GENERAL COMMENTS
Dr. Manoja Kumar Das et al. explored the association 218 between intussusception in Indian children and meteorological parameters. Data for 2161 children aged 2-24 months with primary intussusception from a multisite (20 sites) bidirectional (retrospective and prospective) surveillance in India during July 2010 to September 2017 were used. The association between monthly intussusception cases and meteorological parameters were examined at pooled, regional and site levels using Pearson's and Spearman's rank-order correlation, factorial analysis-ofvariance and Poisson regression or negative binomial regression analyses. They concluded that a seasonal variation in monthly intussusception was observed in India, with a peak in summer months and more evident for three regions. Monthly intussusception cases were positively associated with the monthly mean meteorological parameters, especially temperature. I have some major concerns 1. Data for 2161 children aged 2-24 months with primary intussusception from a multisite (20 sites) bidirectional (retrospective and prospective) surveillance in India during July 2010 to September 2017 were used. And only 2161 children aged 2-24 months with primary intussusception were included. Does these data can be correctly representative the occurrence of intussusception in India? 2. Spearman's rank-order correlation analysis was also conducted to explore whether the monthly intussusception and monthly mean meteorological parameters covary. Factorial analysis-of variance was used to analyse the variations in the intussusception cases during the different seasons (dependent variable: seasonal intussusception cases and independent variable: season and year). Poisson regression or negative binomial regression (for parameters with wide dispersion) analyses was performed to evaluate the relationships between the monthly intussusception cases (dependent variable) and the monthly mean meteorological parameters (independent variables).
1. There have already been many reports of seasonal variation in intussusception in other countries, as the authors have said. The authors have explained the correlation between climate and intussusception through statistical methods, but as the authors are aware of, it is difficult to correct various factors and the causal relationship is not clear. Therefore, the reason for such a tendency is to be supplemented in the discussion. I think that the authors only talked about the association with appendicitis. The reasons described in line 516-521 should have been explained in more detail. As the authors know, changes such as temperature, humidity, sunshine do not cause intussusception directly, but may be due to causes such as infection associated with that seasonal differences. Indeed, many studies on the incidence of intussusception over a long period of time have shown that it decreases with industrialization or improved access to medical care. It is necessary to explain the association with seasonal infection patterns, or to analyze various factors such as regional economic factors and medical accessibility.
2. In 2010, the number of patients is exceptionally small, even considering that the survey was conducted for only six months. What do you think is the reason?
3. Also, the authors said that the large number was one of the strengths of their study, and the studies that the authors referenced showed more than 5000 sample sizes. However, a recent study in South Korea involved 25,000 people. ( The association between monthly intussusception cases and meteorological parameters were examined at pooled, regional and site levels using Pearson's and Spearman's rank-order correlation, factorial analysis-of-variance and Poisson regression or negative binomial regression analyses. They concluded that a seasonal variation in monthly intussusception was observed in India, with a peak in summer months and more evident for three regions. Monthly intussusception cases were positively associated with the monthly mean meteorological parameters, especially temperature. 2.2. Spearman's rank-order correlation analysis was also conducted to explore whether the monthly intussusception and monthly mean meteorological parameters covary. Factorial analysis-of variance was used to analyse the variations in the intussusception cases during the different seasons (dependent variable: seasonal intussusception cases and independent variable: season and year). Poisson regression or negative binomial regression (for parameters with wide dispersion) analyses was performed to evaluate the relationships between the monthly intussusception cases (dependent variable) and the monthly mean meteorological parameters (independent variables). Many methods were used to explore the association between mean meteorological parameters and intussusception. Which statistic methods is the best? The author should discuss it in discussion. Response: We have responded to similar comment from the Editor above. We would like to submit again the following responses for consideration. We used Pearson's correlation coefficient as the primary analysis to document the strength and direction of relationship between the meteorological parameters and intussusception. As secondary analysis and checking for the consistency in the findings, we used Spearman's rank-order correlation analysis to identify strength of association between ranked parameters and intussusception variables. The Pearson's correlation coefficient (r) and Spearman's correlation coefficient (ρ) values for different meteorological parameters were similar, although the r values for temperature, rainfall, wind speed and humidity were higher than the ρ values, but had r value was lower for sunshine than the ρ value at pooled level. The differences between the r and ρ values were relatively small, which comforts and supports the consistent association between the variables. Factorial analysis-of-variance was used to explore the variations in the intussusception cases across seasons and years, which observed significant variation in intussusception cases across seasons at pooled level, for north region, 2-6 months age strata and seasons (summer and autumn). Regression analysis (Poisson or negative binomial regression based on dispersion of the parameters) was done for assess the relationships between the monthly intussusception cases (dependent variable) and the monthly mean meteorological parameters (independent variables). We observed significant association between the intussusception and meteorological parameters (temperature, wind Speed, humidity and sunshine) except rainfall. These associations were consistent with the Pearson's correlation coefficient (r) and Spearman's correlation coefficient values, which further supports the consistency in association.
We have added these in the discussion section. Response: We have revised the tenure to seven years. We have tried to correct the typographical and grammatical mistakes in the manuscript.

Reviewer: 2
Comments to the Author This is very interesting as a study that confirming the occurrence of intussusception according to meteorological parameters in India. Moreover, India is a country with a large area, and because there are differences in distance between South and North, and East and West, the results of regional and climatic differences are meaningful. Response: We thank the learned reviewer for encouraging feedback.
3.1. There have already been many reports of seasonal variation in intussusception in other countries, as the authors have said. The authors have explained the correlation between climate and intussusception through statistical methods, but as the authors are aware of, it is difficult to correct various factors and the causal relationship is not clear. Therefore, the reason for such a tendency is to be supplemented in the discussion. I think that the authors only talked about the association with appendicitis. The reasons described in line 516-521 should have been explained in more detail. As the authors know, changes such as temperature, humidity, sunshine do not cause intussusception directly, but may be due to causes such as infection associated with that seasonal differences. Indeed, many studies on the incidence of intussusception over a long period of time have shown that it decreases with industrialization or improved access to medical care. It is necessary to explain the association with seasonal infection patterns, or to analyze various factors such as regional economic factors and medical accessibility. Response: Seasonal variations in intussusception cases have been reported across several countries (China, Ethiopia, Taiwan, South Africa, and Israel) and also India. Reports from some countries (Switzerland, France, New Zealand, Singapore and Latin American countries including Argentina, Brazil, Chile, Colombia, and Mexico) have reported no definite seasonal variation in the intussusception cases.
The association between the intussusception and meteorological parameters (temperature, sunshine and humidity) and seasonal variations have been documented using statistical tests only from China. We appreciate and acknowledge that it is difficult to establish any causal linkage between the meteorological parameters and intussusception. We are reporting the epidemiological association between the meteorological parameters and intussusception in Indian context and variations across the different regions. Further investigations are needed to establish the causal linkages between the potential aetiological/risk factors (like infectious, dietary, intestinal motility and any other risk factors) and intussusception. Infectious causes have been ascribed as one of the potential cause of intussusception. Also appendicitis has been documented as a lead point trigger for intussusception. Thus, we have given appendicitis as an example to support the exploration.
The potential influence of industrialisation on intussusception has also been added to the discussion. We have added these in the discussion section.
(Manuscript with changes marked; Page 24-25, lines 559-577) 3.2. In 2010, the number of patients is exceptionally small, even considering that the survey was conducted for only six months. What do you think is the reason? Response: The study planned collection of five years data for retrospective surveillance. The Technical Advisory Group (TAG) suggested for initiating the retrospective surveillance from July 2010. The data collection was initiated at the site institutes in January-February 2016. At the institutes usually the medical case record are kept at the hospital record section for minimum of five years and at several institutes, these records are moved to another store after five years. Thus the TAG suggested to start the surveillance from July 2010.
3.3. Also, the authors said that the large number was one of the strengths of their study, and the studies that the authors referenced showed more than 5000 sample sizes. However, a recent study in South Korea involved 25,000 people. ( The study from Korea (near complete record of cases from the country) reported 25023 intussusception cases in children aged 0-18 years of age and the children <2 years of age were 15792 over 10 years. Our study reports findings for 2279 children, aged <2 years, largest from India till date with regional geographical representation. Thus we have mentioned the large sample size as the strength of the study. This network collected data from 20 institutions only. Given the India's size and population this network is very small and far from complete to include all the intussusception cases. The sample included in this study is not complete but representative of the regions.

Reviewer: 3
Comments to the Author This article wants to explore whether the occurrence of Intussusception across India is related to climate. This is a national multi-center study, which shows the intentions of the authors. Although the information collected may not be complete, it should still be representative. Before this paper is accepted, there are some important issues that need to be clarified and improved. Response: We thank the learned reviewer for critical feedback.

Reference #12 and 24 studies about relationship between the intussusception and meteorological parameter had number of studies about 6000 and 14000
How come you mentioned the small sample size in the previous studies in page 13 line 280-282 "The association between meteorological factors and intussusception remains unclear, due to various reasons including shorter study duration, smaller sample size, and no definite statistical methods to explore the association. Response: The references 12 and 24 are studies from China and included 5994 and 13887 intussusception cases, respectively. In the statement referred by reviewer, we intended to highlight the challenges with the reports from several countries that had methodological and sample size related problems to identify the potentially association between intussusception and seasonality and meteorological parameters. We have revised the statement. ((Manuscript with changes marked; Page 13, line 297) Apart from the seasonality, the only report from China reported correlation between intussusception and meteorological parameters using statistical tests and modelling approach (Reference 12 and 24). These articles have reported the association between the intussusception and meteorological parameters (temperature, sunshine and humidity) and seasonal variations. The available reports from United States and Australia reported decline in the intussusception incidence with economic development and industrialisation, which may be due to the improvement in access to healthcare, vaccines/preventive measures, hygiene and sanitation and also reduced infections. Their aims to document the association between intussusception in Indian 217 children and meteorological parameters and examine the regional variations. The association between monthly intussusception cases and meteorological parameters were examined at pooled, regional and site levels using Pearson's ( r) and Spearman's rank-order (ρ) correlation, factorial analysis-of-variance and Poisson regression or negative binomial regression analyses. They concluded that this study documented higher intussusception cases during the summer months with a positive association between number of cases with temperature and sunshine. Also significant positive correlation with the humidity and wind speed was observed. The association was significant for the North and East regions with wider variations in the meteorological parameters across seasons.
Good study, however following needs to be taken care of: 1. 2279 children aged 2-24 months with intussusception were recruited. Out of these, 118 children had past history of intussusception, and were excluded. Why did the author exclud 118 cases in the present study? 2. How did the author define primary intussusception in the present study? 3. Page 24 . the author claimed that some of the intussusception have associate with appendicitis. To the best of my knowledge, very few intussusception in children associated with appendicitis. Is this true? 4. Page 25, "The study had several strengths; large number of intussusception cases over seven years period". large number should be "relative large sample". 5. the design of the present paper is A bidirectional (retrospective and prospective). Are there differences association intussusception in Indian children and meteorological parameters?

GENERAL COMMENTS
The authors have fully revised and supplemented the previously mentioned content. I think the association between Intussusception and seasonality will be different from country to country because there are various factors. Even assuming that the cause of intussusception is only infection, various factors such as the clarity of seasonal changes, economic status of the country, and hygiene will affect. Nevertheless, since intussusception is caused by structural problems that become leading points or many other causes, seasonality is thought to be inevitably different from country to country. I think you've added these parts as well. It seems that the reference I mentioned in the previous review was mentioned in the answer to another reviewer. I think you may add the literature to the reference.

REVIEWER
Chen, Solomon Ditmanson Medical Foundation Chia-Yi Christian Hospital, Department of Pediatrics REVIEW RETURNED 15-Dec-2020

GENERAL COMMENTS
All my previous concerns have been well addressed. Thanks. Their aims to document the association between intussusception in Indian 217 children and meteorological parameters and examine the regional variations. The association between monthly intussusception cases and meteorological parameters were examined at pooled, regional and site levels using Pearson's (r) and Spearman's rank-order (ρ) correlation, factorial analysis-of-variance and Poisson regression or negative binomial regression analyses. They concluded that this study documented higher intussusception cases during the summer months with a positive association between number of cases with temperature and sunshine. Also significant positive correlation with the humidity and wind speed was observed. The association was significant for the North and East regions with wider variations in the meteorological parameters across seasons.

VERSION 2 -AUTHOR RESPONSE
Good study, however following needs to be taken care of: 1. 2279 children aged 2-24 months with intussusception were recruited. Out of these, 118 children had past history of intussusception, and were excluded. Why did the author exclude 118 cases in the present study?
Response: In view of potential differences in the degree of association for the first and recurrent intussusception episodes with the meteorological parameters, we only considered the first intussusception episodes. Thus the 118 children with past history of intussusception were excluded from analysis.