Table 2. Little evidence that FPN Q248H provides protection against severe malaria or invasive bacterial infection.
All P values reflect two-tailed tests. NTS, nontyphoidal Salmonella; RR, relative risk.
| Outcome | All | Q248H* | WT |
RR/OR (95% CI) |
P |
| Malaria susceptibility† | |||||
| Mild malaria | 260/2,550 (10.2%) | 25/232 (10.8%) | 235/2,318 (10.1%) | 1.06 (0.68–1.64) | 0.79 |
| Severe malaria‡ | 5,489/11,982 (45.8%) | 658/1519 (43.3%) | 4,831/10,463(46.1%) | 0.91 (0.81–1.01) | 0.08 |
| (The Gambia§) | 2,419/4,910 (49.3%) | 225/479 (47.0%) | 2,194/4,431 (49.5%) | 0.88 (0.73–1.07) | 0.21 |
| (Malawi§) | 1,023/2,345 (43.6%) | 174/419 (41.5%) | 849/1,926 (44.1%) | 0.89 (0.72–1.10) | 0.28 |
| (Kenya§) | 1,446/2,924 (49.5%) | 191/405 (47.2%) | 1,255/2,519 (49.8%) | 0.98 (0.79–1.22) | 0.87 |
| (Ghana║) | 601/1,803 (33.3%) | 68/216 (31.5%) | 533/1,587 (33.6%) | 0.83 (0.59–1.17) | 0.29 |
| Cerebral malaria‡ | 1,948/7,971 (24.4%) | 249/1043 (23.9%) | 1,699/6,928 (24.5%) | 0.90 (0.77–1.06) | 0.21 |
| (The Gambia§) | 758/3,676 (20.6%) | 60/356 (16.9%) | 698/3,320 (21.0%) | 0.73 (0.55–0.99) | 0.04 |
| (Malawi§) | 644/2,057 (31.3%) | 116/373 (31.1%) | 528/1,684 (31.4%) | 0.95 (0.75–1.22) | 0.71 |
| (Kenya§) | 546/2,238 (24.4%) | 73/314 (23.2%) | 473/1,924 (24.6%) | 1.00 (0.75–1.34) | 0.99 |
| Severe malarial anemia‡ | 735/7,971 (9.2%) | 82/1043 (7.9%) | 653/6,928 (9.4%) | 0.90 (0.71–1.16) | 0.42 |
| (The Gambia§) | 428/3,676 (11.6%) | 43/356 (12.1%) | 385/3,320 (11.6%) | 0.96 (0.68–1.35) | 0.82 |
| (Malawi§) | 91/2,057 (4.4%) | 12/373 (3.2%) | 79/1,684 (4.7%) | 0.69 (0.37–1.29) | 0.24 |
| (Kenya§) | 216/2,238 (9.7%) | 27/314 (8.6%) | 189/1,924 (9.8%) | 0.93 (0.60–1.43) | 0.74 |
| Malaria-related death‡ | 677/4,669 (14.5%) | 86/571 (15.1%) | 591/4,098 (14.4%) | 1.00 (0.78–1.29) | 0.99 |
| (The Gambia§) | 309/2,335 (13.2%) | 27/219 (12.3%) | 282/2,116 (13.3%) | 0.91 (0.59–1.39) | 0.65 |
| (Malawi§) | 200/1,018 (19.6%) | 37/173 (21.4%) | 163/845 (19.3%) | 1.14 (0.76–1.71) | 0.51 |
| (Kenya§) | 168/1,316 (12.8%) | 22/179 (12.3%) | 146/1,137 (12.8%) | 0.94 (0.58–1.52) | 0.80 |
| Bacteremia¶ | 1,536/4,213 (36.5%) | 223/607 (36.7%) | 1,313/3,606 (36.4%) | 1.04 (0.87–1.24) | 0.68 |
| Streptococcus pneumoniae | 426/4,213 (10.1%) | 57/607 (9.4%) | 369/3,606 (10.2%) | 0.93 (0.69–1.26) | 0.65 |
| NTS | 180/4,213 (4.3%) | 31/607 (5.1%) | 149/3,606 (4.1%) | 1.25 (0.83–1.87) | 0.28 |
| Escherichia coli | 151/4,213 (3.6%) | 26/607 (4.3%) | 125/3,606 (3.5%) | 1.27 (0.82–1.97) | 0.29 |
| Haemophilus influenzae | 128/4,213 (3.0%) | 12/607 (2.0%) | 116/3,606 (3.2%) | 0.64 (0.35–1.18) | 0.15 |
| Staphylococcus aureus | 175/4,213 (4.2%) | 20/607 (3.3%) | 155/3,606 (4.3%) | 0.80 (0.50–1.30) | 0.37 |
*Heterozygotes and homozygotes.
†Mild malaria was defined as P. falciparum positive slide measured in community-based cohorts in Uganda, The Gambia, Burkina Faso, and Kenya. Severe malaria was defined as positive for P. falciparum parasites and clinical features of severe malaria (8), including diagnosis of cerebral malaria, severe malarial anemia, and other clinical symptoms.
‡Relative risk, CI, and P value were computed by fixed-effect meta-analysis of estimates from the three case-control cohorts and (where applicable) the Ghanaian trios.
§Association analysis estimated by logistic regression adjusted for the first five principal components. For severe malaria and malaria-related death, results reflect binomial logistic regression of the phenotype compared with controls. For severe malaria subphenotypes, results reflect multinomial logistic regression of cerebral malaria, severe malarial anemia, and other severe malaria cases compared with controls. A dominant mode of effect is assumed.
║Counts reflect numbers of probands (affected children) and parents in 608 Ghanaian trios. Relative risk, CI, and P value are computed using a transmission disequilibrium test. A dominant mode of effect is assumed.
¶Bacteremia was defined as positive blood culture from hospitalized admission in Kilifi County Hospital in Kenya. Association analysis for all-cause bacteremia was estimated by logistic regression adjusted for sex and the first two principal components of genome-wide genotyping data to account for population structure. Pathogen-specific P values and odds ratios were derived by multinomial logistic regression adjusted for sex and population structure.