Supplementary appendix

Transcription

1 Supplementary appendix This appendix formed part of the original submission and has been peer reviewed. We post it as supplied by the authors. Supplement to: Boivin MJ, Okitundu D, Makila-Mabe B, et al. Cognitive and motor performance in Congolese children with konzo during 4 years of follow-up: a longitudinal analysis. Lancet Glob Health 2017; 5: e

2 Web appendix Caldwell Home Observation for the Measurement of the Environment (HOME) and Socioeconomic Status (SES). Following consent and enrollment at baseline, a nurse visited the home of each study child to complete the middle childhood version of the Caldwell Home Observation for the Measurement of the Environment (HOME). 7 This instrument is designed to measure the quality and quantity of stimulation and support available to a child in the home environment. The focus of the assessment is on the child in the environment and the child as a recipient of inputs from objects, events, and transactions occurring in connection with the family surroundings. The nurse also completed an evaluation of the socioeconomic condition of the home This measure scores a dwelling on the basis of type of floor, roofing, wall structure (e.g., mud and wood or cement/brick), cooking facilities, toilet facilities, water source or plumbing, electricity/appliances, and material possessions. It also included measures of the literacy status of the mother and father and their occupation. As adapted for the present study setting, HOME scores can range from 20 to 80 and for the SES from 0 to 10. Kaufman Assessment Battery for Children, 2 nd edition (KABC-II). Neuropsychological evaluation took place in a quiet room near Kahemba Hospital that afforded privacy. The KABC provides a comprehensive evaluation of core neurocognitive performance domains. It was used in our baseline study, and so was the principal measure of cognitive ability for the 2- and follow-up assessments. 4 The principal outcome measure for this test was the Mental Processing Index (MPI) from the Luria model that can be used to evaluate neuropsychological performance on this test. This overall performance measure is a composite of standardized performance on the Sequential Processing (Memory), Simultaneous Processing (visual-spatial analysis and problem solving), Learning, and

3 Planning (reasoning) subtest domains. Delayed Recall (usually about 25 minutes after completing the Learning subtests) was also evaluated. The Knowledge subtests were not administered because they were deemed not appropriate to the present cultural context. Standardized scores (using the KABC-II ASSIST program) based on American agebased norms were used for statistical evaluation in the present study tables (Tables 1 5). In order to obtain standardized performance measures for the KABC-II global and composite scores (with a possible range of 40 to 160 depending on the scale), it is necessary to use American-based norms to obtain scaled scores for the raw scores for each subtest. This is because normative data are not available for DR Congolese children (urban or rural which is an important distinction in evaluating neurocognitive outcomes in African children). Also, no non-exposure reference group (not at-risk for konzo from poorly processed cassava, from outside the outbreak zone) were available at 2- and follow-up as a means of allowing for us to use a local control or reference group for obtaining age-adjustment standardization for the KABC-II test. Such standardization for subtests and global domains is necessary in order to obtain an overall composite or total score (e.g., Mental Processing Index (MPI); Nonverbal Index (NVI)) a composite score for those subtests not dependent on spoken language. Bruininks/Oseretsky Test of Motor Proficiency (2 nd Edition; BOT-2). This is a comprehensive test of motor proficiency used in the baseline study. 4 Composite scores include fine manual control, manual coordination, body coordination, strength and agility, and a total composite score. American based norms were used to standardize the measures on the basis of age and gender (with standard scores generally ranging from 20 to 80 depending on the measure). Because this was the measure used to assess motor proficiency, and had a very significant positive correlation to the neurological impairment index for upper-motor neuron function in the neurological exam for children with konzo in the baseline study, 4 it

4 was also a principal outcome for performance of children with konzo over time for our present follow-up assessments. Both the KABC-II and BOT-2 were used in this follow-up assessment because they also proved sensitive to metabolic biomarkers of cyanide exposure from poorly processed cassava in the baseline assessment Evaluation of the potential mediational role of WAZ for BOT-2 and KABC-II principal outcomes. We repeated the GLME analyses presented in the analyses of the BOT-2 (Tables 2 and 3) and KABC-II principal outcomes (Tables 4 and 5), this time adding WAZ at baseline as an additional covariate. When adding WAZ as a covariate, the findings presented in Tables 2 through 5 in terms of the differences between children with and without konzo did not change in any appreciable way. Therefore, wasting did not serve as a mediator in our GLME models for the effects of konzo on our principal BOT-2 and KABC-II outcomes at baseline, two-, and four-year assessment for either the boys or girls. However, in terms of its covariate main effect across both konzo groups and across all three time points (baseline, two year, four year), WAZ was significantly related with BOT-2 Fine Manual motor proficiency scores for both sexes (Boys: Beta=0 034, SE=0 013, t(108)= 2 52, p= ; Girls: Beta=0 044, SE=0 022, t(87)= 2 03, p= ). For the girls, WAZ was also a significant predictor of BOT-2 Manual Coordination (Beta=0 046, SE=0 022, t(87)= 2 08, p= ), Strength/Agility (Beta=0 044, SE=0 020, t(87)= 2 21, p= ), and Total Motor Score (Beta=0 046, SE=0 018, t(87)= 2 51, p= ). Only the relationship between WAZ and BOT-2 Total Motor score for the girls remained statistically significant after a Bonferroni adjustment for the number of simultaneous comparisons for the BOT-2 principal outcomes. In terms of the WAZ covariate main effect for any of the KABC-II principal outcomes across both konzo groups and across all three time points (baseline, two year, four year), WAZ was significantly correlated with KABC-II Simultaneous Processing (visual-spatial analysis and

5 problem solving) for boys (Boys: Beta=0 026, SE=0 012, t(108)= 2 18, p= ). No other KABC-II outcomes for boys or girls were significantly related to WAZ. Evaluation of the potential mediational role of HAZ for BOT-2 and KABC-II principal outcomes. We then repeated the GLME analyses presented in the analyses of the BOT-2 (Tables 2 and 3) and KABC-II principal outcomes (Tables 4 and 5), this time adding HAZ at baseline as an additional covariate. Controlling for HAZ at baseline did not change the results for the differences between kids with and without konzo for the BOT-2 outcomes for both the boys (Table 2) and girls (Table 3). This indicates that stunting does not mediate the between-group differences between children with and without konzo on motor proficiency. This is despite the fact that HAZ itself was significantly related to all BOT-2 outcomes. The only exception to this is for the BOT-2 subscale of Body Coordination for the boys (Beta=0 026, SE=0 014, t(108)= 1 82, p= 0 07) and girls (Beta=0 016, SE=0 011, t(87)= 1 47, p= 0 15). When controlling for HAZ, the BOT-2 Fine Manual subscale differences between children with and without konzo changed to insignificant at baseline (P=0 29) and significant at two-year (p<0 01) and four-year (p<0 01) follow-up for the boys. Controlling for HAZ in the analysis of the KABC-II resulted in the attenuation of the few significant between-group differences for the boys in Table 4. Differences for the KABC-II Learning domain between boys with and without konzo becomes insignificant when controlling for HAZ. Likewise, KBC-II MPI for the boys at two-year follow-up becomes insignificant when controlling for HAZ. This suggests a mediational role of stunting when comparing boys with and without konzo in these two KABC-II domains. The significant difference in the Planning subscale of the KABC-II between boys with and without konzo at baseline (p=0 0056, Table 4) remained significant when controlling for HAZ (p=0 0218), suggesting that HAZ only partially mediated this between-group

6 difference. Among boys, HAZ as a covariate across groups and time of assessment was significantly related to all KABC-II global domain scores except Planning. For the girls, HAZ was not related to any of the KABC-II global domain scores. Also, controlling for HAZ in the GLME models presented in Table 5 did not change the lack of between-group konzo/nonkonzo KABC-II differences at any of the assessment time points (baseline, 2 year, 4 year) for the girls. Supplemental Table. Baseline characteristics by and by follow-up completion ( versus drop-outs) and konzo status. Data are means followed by standard deviations in parentheses. P values are for the Kruskal-Wallis one-way ANOVA comparison of versus drop outs. Follow-up Characteristic N=76 Konzo drop-outs N=47 p value N=36 Nonkonzo drop-outs N=46 P value HOME score (7 17) (8 92) (11 84) (8 10) 0 12 KABC MPI (7 74) (8 62) (9 92) (8 61) 0 08 BOT-2 Total (6 32) (5 51) (6 42) (8 16) 0 33 WAZ CDC (0 97) (0 96) (1 73) (1 27) 0 10 HAZ CDC (1 35) (1 52) (1 22) (2 15) 0 12 Follow-up Characteristic N=55 Konzo drop-outs, N=68 p value N=33 Nonkonzo drop-outs N=49 p value HOME score (7 98) (7 96) (10 83) (9 18) 0 08 KABC MPI (7 31) (8 83) (8 68) (9 67) 0 18 BOT-2 Total (6 60) (5 50) (8 34) (6 86) 0 80 WAZ CDC (1 06) (0 90) (2 04) (0 93) 0 13 HAZ CDC (1 32) (1 50) (2 29) (1 34) 0 08

7