Recurrence of WHO-defined fast breathing pneumonia among infants, its occurrence and predictors in Pakistan: a nested case–control analysis

Objectives Studies in low-income and middle-income countries have shown an adverse association between environmental exposures including poverty. There is little literature from South Asia. We aimed to test the associations between housing, indoor air pollution and children’s respiratory health and recurrent fast breathing pneumonia in a poor urban setting in Pakistan. Setting Primary health centres in a periurban slum in Karachi, Pakistan. Methods Nested matched case–control study within a non-inferiority randomised controlled trial of fast breathing pneumonia (Randomised Trial of Amoxicillin vs Placebo for Pneumonia (RETAPP)) in periurban slums of Karachi, Pakistan. Cases were children aged 2–60 months enrolled in RETAPP with fast breathing pneumonia who presented again with fast breathing between 8 weeks and 12 months after full recovery. Controls, selected in a 2:1 ratio, were age-matched participants who did not represent. Multivariable conditional logistic regression analysis was undertaken to explore associations with potentially modifiable environmental predictors including housing type, indoor air quality, exposure to tobacco smoke, outdoor pollution, household crowding, water and sanitation quality, nutritional status, immunisation completeness, breast feeding and airways hyperactivity. Results Fast breathing recurred in 151 (3.7%) of children out of the total (4003) enrolled in the trial. Poor-quality housing of either katcha or mixed type strongly predicted recurrence with adjusted matched ORs 2.43 (95% CI 1.02 to 5.80) and 2.44 (1.11 to 5.38), respectively. Poor air quality, cooking fuel, inadequate ventilation, nutritional status, water, sanitation and hygiene (WASH) index, wheeze at first presentation and group of initial trial assignment were not independently predictive of recurrence. Conclusion Poor-quality housing independently predicted recurrence of fast breathing pneumonia. Trial registration number NCT02372461

MAIN TEXT • Background: I think the introduction is fundamentally sound, updated, reflecting the importance of the recurrent pneumonia and the paucity of evidence, especially from developing countries. • Methods: Just curious to know if the case definition was supported by radiological findings, or it was just based on recurrent episode of isolated fast breathing between eight weeks to one year of the initial episode. Was any assessment made based on the season of the year? Were there more cases of RP in winters or rainy season? Was any information collected on underlying conditions such as, family history of asthma, congenital cardiac defects, immunological abnormalities, premature delivery, and neuromuscular disorders etc.? Was temporary migration in the slum taken into account? Could the authors explain what is meant by an open and closed stove? Does closed stove refer to stove using natural gas, or it is a conventional stove, but not an open fire? In Table 1 (page 11), the proportion of cases and controls using an open stove is quite high, 94% and 89% respectively, but the solid fuel use is only 24% and 12% respectively. I don't understand the mismatch. Could this be clarified? As the parent trial has been administering oral amoxicillin to the children with fast breathing, were there any children who didn't complete the 3 day course? Were these children part of the nested case-control analysis? Sample size calculation is not clear enough. Why was solid fuel use chosen as the main predictor for RP? Was there any specific hypotheses?
Overall, the methods section needs clarification. • Results: The results mention that only solid fuel use independently predicted recurrent fast breathing, however, in the absence (in case) of assessment of other potential predictors as mentioned above that could have important implications, the results must be interpreted carefully. The authors have transparently reported the limitations of the study, including small sample size, few cases due to lack of database linkage, underestimation of recurrent cases due to presentation to other health facilities, and inability to quantify solid fuel exposure.

REVIEWER
Thi Kim Phuong Nguyen Da Nang hospital for Women and Children, Viet Nam REVIEW RETURNED 17-Sep-2017

GENERAL COMMENTS
This is an interesting study, which provided the audience some risk factors of recurrent pneumonia. Some risk factors of child pneumonia are recognized by WHO but unsure their role in recurrent pneumonia. The study showed a strong correlation between indoor air pollution exposure with increase the risk and completed vaccination as a protected factor. It will be more interesting if the authors considered 2 other important risk factors as exclusive breastfeeding and co-morbid conditions in these recurrent cases.

Recurrence of World Health Organization defined Fast Breathing Pneumonia among Infants, its Occurrence and Predictors in Pakistan-A Nested Case Control Analysis.
Comments This is an interesting study, which provided the audience with suggested predictors of recurrent pneumonia. Some risk factors of child pneumonia are recognized by WHO but unsure their role in recurrent pneumonia. The study showed a strong correlation between indoor air pollution exposure with increase the risk and completed vaccination as a protected factor. It will be more interesting if the authors considered 2 other important risk factors as exclusive breastfeeding and co-morbid conditions in these recurrent cases. I just have minor suggestions Introduction section Page 5: It that possible if the authors just use pneumonia or ALRI instead of them both in the text? This may sometimes be confused. Line 22, page 5: in revised WHO pneumonia classification 2014, "non-severe pneumonia" = fast breathing +/-chest indrawing; "severe pneumonia" = fast breathing + any danger signs. Please check your text to make this clear and consistent.
Method section Some points that may affect your results and need to be considered 1. Did these cases visit and receive any treatment from private doctor, pharmacist, traditional healer before admitting to the PHC? 2.
Did they go to child care or had any siblings? If that, was there any sick children (with ARI symptoms) at school or home? 3.
History of TB contact, etc… Result section I would prefer a confidence interval belongs to each OR in table 1 Discussion section Consider these points in method section that I suggested above Please delete 1 of 2 "from" in line 11, page 15 Reference You can find the update number of child pneumonia disease burden from this article Li Liu, Shefali Oza, Dan Hogan, Yue Chu, Jamie Perin, Jun Zhu, et al. Global, regional, and national causes of under-5 mortality in 2000-15: an updated systematic analysis with implications for the Sustainable Development Goals. Lancet. 2016;16:31593-8.

REVIEWER
Carina King University College London, UK REVIEW RETURNED 18-Sep-2017

GENERAL COMMENTS
This is an important topic, and one which definitely needs more investigation, so this is a relevant paper. It is well presented and clear, and I only have minor comments for clarification and two requests for a bit more analysis and data to be presented. A rigorous proof-reading is also needed.

Strengths & Limitations: -'A priori'
Introduction: -In the description of ALRI you discuss chest in-drawing pneumonia as severe. This needs to be made a bit clearer, as the revised IMCI guidelines have chest in-drawing pneumonia being treated as an outpatient (pneumonia rather than severe pneumonia). Needs a clarification about which guidelines you are using.
Methods: -Study design and setting: Was chest in-drawing included in the danger signs? It might be useful to list the danger signs just to clarify for the reader (e.g. wheeze?).
-Study design and setting: Who conducted the follow-ups, were they home or facility based? -Study population: it would be useful to know a bit more about where geographically the study was conducted.
-Data collection: how was vaccine status recorded, from documented records, or caregiver recall? And was wheeze assessed by the study physician, or was this care-giver reported history?
Results: -Were the randomly selected controls representative of all the potential controls? A supplementary I think the introduction is fundamentally sound, updated, reflecting the importance of the recurrent pneumonia and the paucity of evidence, especially from developing countries.

Response
Thank you

Methods
Just curious to know if the case definition was supported by radiological findings, or it was just based on recurrent episode of isolated fast breathing between eight weeks to one year of the initial episode.

Response
In keeping with the WHO pragmatic syndromal management of fast breathing pneumonia, no x rays were undertaken in children without complications

Response
We have simplified our predictors to include only solid (wood, coal and animal dung) and gas 8. As the parent trial has been administering oral amoxicillin to the children with fast breathing, were there any children who didn't complete the 3 day course? Were these children part of the nested case-control analysis?

Response
No. Only children who completed treatment in the trial were eligible 9. Sample size calculation is not clear enough. Why was solid fuel use chosen as the main predictor for RP? Was there any specific hypotheses?
Overall, the methods section needs clarification.

Response
Thanks for this helpful comment. The study was in part exploratory and hypothesis generating, but, to inform sample size we made an estimate based on a posited effect size (OR 2.5) of public health importance for the composite indoor air pollution variable on recurrent pneumonia 10 Results The results mention that only solid fuel use independently predicted recurrent fast breathing, however, in the absence (in case) of assessment of other potential predictors as mentioned above that could have important implications, the results must be interpreted carefully.
The authors have transparently reported the limitations of the study, including small sample size, few cases due to lack of database linkage, underestimation of recurrent cases due to presentation to other health facilities, and inability to quantify solid fuel exposure.

Response
Thank you. We have been careful to avoid over interpretation of the results Reviewer: 2 1. This is an interesting study, which provided the audience some risk factors of recurrent pneumonia. Some risk factors of child pneumonia are recognized by WHO but unsure their role in recurrent pneumonia. The study showed a strong correlation between indoor air pollution exposure with increase the risk and completed vaccination as a protected factor. It will be more interesting if the authors considered 2 other important risk factors as exclusive breastfeeding and co-morbid conditions in these recurrent cases.

Response
Thank you. Breast feeding was not predictive or protective in the univariable model so was not included in the multivariable analysis. Breast feeding is however, the norm so differences might be harder to detect … 2. It that possible if the authors just use pneumonia or ALRI instead of them both in the text? This may sometimes be confused.

Response
Thank you. We have now used the term pneumonia throughout 3. In the revised WHO pneumonia classification 2014, "nonsevere pneumonia" = fast breathing +/chest indrawing; "severe pneumonia" = fast breathing + any danger signs. Please check your text to make this clear and consistent.

Response
Thank you. We have reworded this section 4. Did these cases visit and receive any treatment from private doctor, pharmacist, traditional healer before admitting to the PHC?

Response
We specifically enquired about additional treatment. Any children that had received antibiotics for any indication within 48 hours of presentation were excluded 5. Did they go to child care. Was sibling number addressed. Were there sick children at home. History

Response
It is traditional for all children in this community to be cared for at home We do not have data on concurrent illness in siblings Sibling number was included as a co-variate in the regression model. It was not significant Known tuberculosis was an exclusion criterium 6. I would prefer a confidence interval belongs to each OR in table 1

Response
These have now been included 7. Delete 'from' line 11, page 15

Response
Deleted 8. Additional reference for child pneumonia disease burden from this article

Response
Thank you. We have included the Liu, Lancet 2016 reference as suggested In the description of ALRI you discuss chest in-drawing pneumonia as severe. This needs to be made a bit clearer, as the revised IMCI guidelines have chest in-drawing pneumonia being treated as an outpatient (pneumonia rather than severe pneumonia). Needs a clarification about which guidelines you are using. -Study design and setting: Who conducted the follow-ups, were they home or facility based?

Response
Thank you In the RETAPP trial follow ups which included directly observed treatment and a clinical assessment were conducted by physicians in the mornings and community health workers in the evenings -Study population: it would be useful to know a bit more about where geographically the study was conducted.

Response
Detail in methods -Data collection: how was vaccine status recorded, from documented records, or caregiver recall?
And was wheeze assessed by the study physician, or was this care-giver reported history? Thank you for addressing my previous comments -I think the methods are now clearer, and using a matched analysis has improved the results. There are still a few minor grammatical errors, and I think one of the < or > are the wrong way around. In the results, poor housing is described as katcha and mixed -but these need to be explained. It would be better to express as "associations with" rather than "predictive of" in the results.

Response
Thank you for the generous comments Amended as suggested