Contribution of malaria and sickle cell disease to anaemia among children aged 6–59 months in Nigeria: a cross-sectional study using data from the 2018 Demographic and Health Survey

Objectives To estimate the fraction of anaemia attributable to malaria and sickle cell disease (SCD) among children aged 6–59 months in Nigeria. Design Cross-sectional analysis of data from Nigeria’s 2018 Demographic and Health Survey (DHS). Setting Nigeria. Participants 11 536 children aged 6–59 months from randomly selected households were eligible for participation, of whom 11 142 had complete and valid biomarker data required for this analysis. Maternal education data were available from 10 305 of these children. Primary outcome measure Haemoglobin concentration. Results We found that 70.6% (95% CI: 62.7% to 78.5%) of severe anaemia was attributable to malaria compared with 12.4% (95% CI: 11.1% to 13.7%) of mild-to-severe and 29.6% (95% CI: 29.6% to 31.8%) of moderate-to-severe anaemia and that SCD contributed 0.6% (95% CI: 0.4% to 0.9%), 1.3% (95% CI: 1.0% to 1.7%) and 10.6% (95% CI: 6.7% to 14.9%) mild-to-severe, moderate-to-severe and severe anaemia, respectively. Sickle trait was protective against anaemia and was associated with higher haemoglobin concentration compared with children with normal haemoglobin (HbAA) among malaria-positive but not malaria-negative children. Conclusions This approach used offers a new tool to estimate the contribution of malaria to anaemia in many settings using widely available DHS data. The fraction of anaemia among young children in Nigeria attributable to malaria and SCD is higher at more severe levels of anaemia. Prevention of malaria and SCD and timely treatment of affected individuals would reduce cases of severe anaemia.

Specific comments: Table 1 -Is the final column numbers or percents? Why is it formatted differently from the other columns? Usually I would expect to see the total column at the end or beginning. Results: Line 27: "Severe anaemia was more prevalent among younger children except for those with HbSS"there is no statistical test to support this and visually all lines except that for HbSS appear flat. Table 2 -This table is a little disappointing given the rich breakdown  in Table 1, why are you only looking at moderate to severe anemia here when using so many other categories in all other analysis? What about HbAA and other categories? What are these p-values from? Chi-square tests? Comparison groups are not obvious for all categories (i.e. maternal education). Please include the N with the percents and include SD with means. Table 3 -When you have odds ratios of greater than 1000 it indicates there is a problem with your modelthis is clearly a result of having no children in the not anemic categorythe model is not valid here as it requires positivity (a non-zero probability of having an outcome level) so you can't include results from it. Table 3 -I think it would be more interesting to use the comparison  group of HbSS negative for the models for HbSSthis would give you a separate odds ratio for the effect of having sickle cell (HbSS negative compared to HbAA negative) and for the effect of having malaria among those with sickle cell (HbSS positive compare to HbSS negative)and do the same for all comparisons. Page 6, lines 48-51in order to say that HbAS was protective you need to provide a statistical comparison, just looking at and comparing the odds ratios is insufficient. Same problem with page 7, lines 23-25 and page 7 lines 40-41, page 8, line 37-38 -If this is one of your main findings, please do some actual statistical testsstatistical software should easily produce odds ratios with specific comparisons of interest Page 7, lines 30-32this would seem to suggest that SS trait is not important for determining the influence of malaria on anemia, again making a change in references for tables 3 and 4 seem useful

REVIEWER
Chunda This manuscript describes the relationship between sickle cell disease, malaria, and anemia in a cross-sectional study of Nigerian children under the age of 5. There is a rich dataset underlying their analyses and the use of attributable fractions is a strength of this paper. However, much of the data is not presented clearly which makes interpretation slightly difficult.
We decided to remove all results related to "semi-severe" anemia, which is not a standard classification and made the tables and text confusing.
General comment -The survey used only RDT to measure malaria. Older children in particular may carry parasites below RDT detectable thresholds for some time, potentially leading to misclassification of some RDT-individuals as malaria negative. The effect of this should be mentioned in the discussion and authors should be more careful to distinguish between RDT negative and malaria negative.
Thank you for this observation. We added a reference to the differences in sensitivity by age and region and added a new point to the "strengths and limitations" section.
Specific comments: Table 1 -Is the final column numbers or percents? Why is it formatted differently from the other columns? Usually I would expect to see the total column at the end or beginning.
We eliminated this confusing column on "semi-severe" anemia. In the new table, the final column is the total number of children per row.
Results: Line 27: "Severe anaemia was more prevalent among younger children except for those with HbSS"there is no statistical test to support this and visually all lines except that for HbSS appear flat.
We agree. The important feature is the sharp increase in severe anemia with age among those with HbSS, and we changed the text to reflect this. We added comparisons for mild and severe anemia to Table 3. We now describe the source of pvalues in the caption. We clarified some of the wording, and added N for the maternal education category, which is missing some data so has a lower N. Table 3 -When you have odds ratios of greater than 1000 it indicates there is a problem with your modelthis is clearly a result of having no children in the not anemic categorythe model is not valid here as it requires positivity (a non-zero probability of having an outcome level) so you can't include results from it.
You are correct that having no children in the "not anemic" category prevents us from estimating these odds ratios. We found that the other odds ratios in the same model were not changed when these problematic categories were excluded (we re-fit the model excluding malaria-positive children with HbSS or HbSC). We changed the last column of Table 3 to summarize the odds of severe anemia (instead of semi-severe), and we encountered the same issue because there are no severely anemic children who are malaria-negative with HbAC or HbSC. We added material describing this to the Methods. Table 3 -I think it would be more interesting to use the comparison group of HbSS negative for the models for HbSSthis would give you a separate odds ratio for the effect of having sickle cell (HbSS negative compared to HbAA negative) and for the effect of having malaria among those with sickle cell (HbSS positive compare to HbSS negative)and do the same for all comparisons.
HbSS/+ has a small number of subjects (24.5) and all of them have at least mild anemia, so we would not be able to fit the model for mild-to-severe anemia using this reference group. We chose HbAA/-(with 5564 children it is the largest category) as the reference group.
Page 6, lines 48-51in order to say that HbAS was protective you need to provide a statistical comparison, just looking at and comparing the odds ratios is insufficient. Same problem with page 7, lines 23-25 and page 7 lines 40-41, page 8, line 37-38 -If this is one of your main findings, please do some actual statistical testsstatistical software should easily produce odds ratios with specific comparisons of interest Thank you for keeping us honest here. We now use RDT-positive children as the reference to compute the protection of HbAS among RDT-positive children (page 6), using RDT-positive children with HbAA as a reference. We added two supplemental tables to summarize these results. We make a note of this in Methods. This made redundant the paragraph where we estimated the anemia prevalence if children with HbAS had HbAA instead, so we removed it.
Page 7, lines 30-32this would seem to suggest that SS trait is not important for determining the influence of malaria on anemia, again making a change in references for tables 3 and 4 seem useful We added two supplemental tables that replicate Tables 3 and 4 but use RDT+ children as the reference.