A flexible approach to measurement error correction in case-control studies

SUMMARY: We investigate the use of prospective likelihood methods to analyze retrospective case-control data where some of the covariates are measured with error. We show that prospective methods can be applied and the case-control sampling scheme can be ignored if one adequately models the distribution of the error-prone covariates in the case-control sampling scheme. Indeed, subject to this, the prospective likelihood methods result in consistent estimates and information standard errors are asymptotically correct. However, the distribution of such covariates is not the same in the population and under case-control sampling, dictating the need to model the distribution flexibly. In this article, we illustrate the general principle by modeling the distribution of the continuous error-prone covariates using the skewnormal distribution. The performance of the method is evaluated through simulation studies, which show satisfactory results in terms of bias and coverage. Finally, the method is applied to the analysis of two data sets which refer, respectively, to a cholesterol study and a study on breast cancer.     

Novel two-phase sampling designs for studying binary outcomes

In biomedical cohort studies for assessing the association between an outcome variable and a set of covariates, usually, some covariates can only be measured on a subgroup of study subjects. An important design question is—which subjects to select into the subgroup to increase statistical efficiency. When the outcome is binary, one may adopt a case-control sampling design or a balanced case-control design where cases and controls are further matched on a small number of complete discrete covariates. While the latter achieves success in estimating odds ratio (OR) parameters for the matching covariates, similar two-phase design options have not been explored for the remaining covariates, especially the incompletely collected ones. This is of great importance in studies where the covariates of interest cannot be completely collected. To this end, assuming that an external model is available to relate the outcome and complete covariates, we propose a novel sampling scheme that oversamples cases and controls with worse goodness-of-fit based on the external model and further matches them on complete covariates similarly to the balanced design. We develop a pseudolikelihood method for estimating OR parameters. Through simulation studies and explorations in a real-cohort study, we find that our design generally leads to reduced asymptotic variances of the OR estimates and the reduction for the matching covariates is comparable to that of the balanced design.     

Common variants of ROCKs and the risk of hypertension,and stroke: Two case-control studies and a follow-up study in Chinese Han population

The Rho kinases (ROCKs) are recognized as a critical regulator of vascular functions in cardiovascular disorders. It is crucial to illustrate the association of ROCKs genetic variation and hypertension and/or stroke events. Herein we aimed at investigating the association of ROCK1 and ROCK2 with hypertension and stroke in Chinese Han population. Seven tagSNPs at ROCK1 and ROCK2 were genotyped in a community-based case-control study consisting of 2012 hypertension cases and 2210 normotensive controls and 4128 subjects were further followed up. In stroke case-control study, 1471 ischemic stroke (IS) inpatients and 607 hemorrhagic stroke (HS) inpatients were collected, and 2443 age-matched controls were selected from the follow-up population. Risks were estimated as odds ratio (OR) and hazard ratio (HR) by logistic and Cox regression. The community-based case-control study didn't identify any significant tagSNPs associated with hypertension even after adjustment for covariates. The follow-up analysis showed that rs1481280 of ROCK1 significantly associated with incident hypertension (HR = 1.130, P = 0.048) after adjusting for covariates. rs7589629 and rs978906 of ROCK2 were significantly associated with incident IS (HR = 1.373, P = 0.004; HR = 1.284, P = 0.026) respectively. In stroke case-control study, rs288980, rs1481280 and rs7237677 were significantly associated with IS and the adjusted ORs (P values) of additive model were 0.879 (0.010), 0.895 (0.036) and 0.857 (0.002) respectively. Furthermore, rs288980, rs7237677 and rs978906 were significantly associated with HS and the adjusted ORs (P values) of additive model were 0.857 (0.025), 0.848 (0.018) and 0.856 (0.027) respectively. Our findings suggest that ROCK1 and ROCK2 contribute to the genetic susceptibility of hypertension and stroke.     

Reticulate evolution in the apogamous Dryopteris varia complex (Dryopteridaceae,subg. Erythrovariae,sect. Variae) and its related sexual species in Japan

Apogamous fern species are often difficult to distinguish from related species because of their continuous morphological variations. To clarify the genetic relationships among the members of the Dryopteris varia complex, we analyzed the nucleotide sequences of the plastid gene rbcL and the nuclear gene PgiC. We also analyzed the diploid sexual species D. caudipinna and D. chinensis, which have not been included in the complex, but were recently shown to be closely related to the complex in a molecular phylogenetic study. The PgiC sequences of the diploid sexual species, D. varia, D. saxifraga, D. sp. ‘protobissetiana’ (undescribed diploid sexual species), D. caudipinna, and D. chinensis, were well differentiated and hence designated A, B, C, D, and E, respectively. Thus, the PgiC constitution of apogamous species in the complex was as follows: D. bissetiana, B + C; D. kobayashii, B + C + E); D. pacifica, A + C, A + B + C, or A + C + D; D. sacrosancta, A + C + E; and D. saxifragivaria, B + C. These results suggest that these apogamous species are formed by hybridizations of species including not only the three diploid sexual species of the D. varia complex (A, B, and C) but also the two diploid sexual species D. caudipinna (D) and D. chinensis (E), which do not belong to the complex.     

Hazard rate models with covariates.

Many problems, particularly in medical research, concern the relationship between certain covariates and the time to occurrence of an event. The hazard or failure rate function provides a conceptually simple representation of time to occurrence data that readily adapts to include such generalizations as competing risks and covariates that vary with time. Two partially parametric models for the hazard function are considered. These are the proportional hazards model of Cox (1972) and the class of log-linear or accelerated failure time models. A synthesis of the literature on estimation from these models under prospective sampling indicates that, although important advances have occurred during the past decade, further effort is warranted on such topics as distribution theory, tests of fit, robustness, and the full utilization of a methodology that permits non-standard features. It is further argued that a good deal of fruitful research could be done on applying the same models under a variety of other sampling schemes. A discussion of estimation from case-control studies illustrates this point.     

Association of FATP1 gene polymorphisms with chicken carcass traits in Chinese meat-type quality chicken populations

In this study, we aimed to detect the single nucleotide polymorphisms (SNPs) of the chicken FATP1 gene and discern the potential association between FATP1 SNPs and chicken carcass traits. A total of 620 meat-type quality chickens from six commercial pure lines (S01, S02, S03, S05, S06 and D99) and two cross lines (S05 × S01 and S06 × S01) were screened by using the single-strand conformational polymorphism analysis (SSCP) and DNA sequencing. Five SNPs [g.49360G > A, g.48195G > A, g.46847A > G, g.46818A > G, and g.46555A > G] were identified in chicken FATP1 gene. SNP g.46818 A > G was a rare variant and was not considered in the subsequent analysis. Sixteen haplotypes were reconstructed on the basis of the other four SNPs. The linear regression model analysis indicated that there were significant associations of certain diplotypes with part of carcass traits, such as live weight (LW), carcass weight (CW), and semi-eviscerated weight (SEW) (P < 0.05). In particular, diplotype H2H4 had a negative effect on LW, CW, SEW, and abdominal fat weight (AW); diplotype H6H10 had the highest reducing effect on subcutaneous fat thickness (SFT). Our results suggested that FATP1gene polymorphisms were associated with chicken carcass traits or was linked with the major gene. The SNPs in this gene may be utilized as potential markers for marker-assisted selection (MAS) during chicken breeding.     

Linkage disequilibrium in two European F<Subscript>2</Subscript> flint maize populations under modified recurrent full-sib selection

According to quantitative genetic theory, linkage disequilibrium (LD) can hamper the short- and long-term selection response in recurrent selection (RS) programs. We analyzed LD in two European flint maize populations, KW1265 × D146 (A × B) and D145 × KW1292 (C × D), under modified recurrent full-sib selection. Our objectives were to investigate (1) the decay of initial parental LD present in F₂ populations by three generations of intermating, (2) the generation of new LD in four (A × B) and seven (C × D) selection cycles, and (3) the relationship between LD changes and estimates of the additive genetic variance. We analyzed the F₂ and the intermated populations as well as all selection cycles with 104 (A × B) and 101 (C × D) simple sequence repeat (SSR) markers with a uniform coverage of the entire maize genome. The LD coefficient D and the composite LD measure Δ were estimated and significance tests for LD were performed. LD was reduced by intermating as expected from theory. A directional generation of negative LD between favorable alleles could not be observed during the selection cycles. However, considerable undirectional changes in D were observed, which we attributed to genetic sampling due to the finite population size used for recombination. Consequently, a long-term reduction of the additive genetic variance due to negative LD was not observed. Our experimental results support the hypothesis that in practical RS programs with maize, LD generated by selection is not a limiting factor for obtaining a high selection response.     

The polymorphisms of LYRM1 gene and their association with body measurement and ultrasound traits of Qinchuan cattle

Body measurement and meat quality traits which play important roles in the assessment of productivity and economy in cattle were influenced by genes and environmental factors. Latest studies showed that LYR motif containing 1 (LYRM1) may be involved in influencing fatness deposition in animals. The objective of this study was to detect bovine LYRM1 gene polymorphism and analyze its association with body measurement and meat quality traits of cattle. Blood samples were taken from a total of 404 Qinchuan cattle aged from 18–24 months. Created restriction site-polymerase chain reaction-restriction fragment length polymorphism (CRS-PCR–RFLP) and DNA sequencing were used to find out LYRM1 single polymorphism nucleotide (SNPs). Sequence analysis of LYRM1 gene revealed two SNPs (g.165 C > A, g.193 A > G) in 3′ untranslated region (3′UTR) of exon 3. And g.165 C > A showed two genotypes namely AC and CC while g.193 A > G showed three genotypes: AA, AG and GG. Analysis results showed that there were significant associations between polymorphism of these two and body measurement and meat quality traits in Qinchuan cattle population. Based on the results obtained from this study, it is inferred that LYRM1 gene may have potential effects on body measurement and meat quality traits in Qinchuan cattle population and could be used for marker-assisted selection.     

Betaine–homocysteine methyltransferase 742G>A polymorphism and risk of down syndrome offspring in a Brazilian population

Down syndrome (DS) is the most common form of mental retardation of genetic etiology. Several polymorphisms in genes involved with the folic acid cycle have been associated to the risk of bearing a DS child; however, the results are controversial. Betaine-homocysteine methyltransferase (BHMT) is a key enzyme of folate pathway, and catalyzes the remethylation of homocysteine into methionine. Recent studies suggest that the polymorphism BHMT 742G>A may be associated with a decreased risk of having a DS child. We herein investigate the association of this polymorphism with the occurrence of DS in a Brazilian population. We have genotyped 94 mothers of DS infants (DSM) and 134 control mothers (CM) for this polymorphism through PCR–RFLP, and found significant differences for both BHMT 742G>A genotype (P = 0.04) and allele (P = 0.03) frequencies between DSM and CM. The observed genotypic frequencies were GG = 0.45; GA = 0.45 and AA = 0.10 in CM, and GG = 0.54; GA = 0.38 and AA = 0.02 in DSM. Allelic frequencies were G = 0.68 and A = 0.32 in CM and G = 0.78 and A = 0.22 in DSM. The presence of the mutant BHMT 742 A allele decreases 40 % the risk of bearing a DS child (OR = 0.61; 95 % CI: 0.40–0.93; P = 0.03), and the risk is diminished up to >80 % in association with the homozygous genotype (OR = 0.17; 95 % CI: 0.04–0.80; P = 0.01). Our results indicate that women harboring the single nucleotide polymorphism BHMT 742G>A have a decreased risk of a DS pregnancy, and further studies are necessary to confirm this protective effect.     

Relationships of LDLR genetic polymorphisms with cerebral infarction: a meta-analysis

This meta-analysis was undertaken to identify the relationships between genetic polymorphisms in the LDLR gene and the risk of cerebral infarction. The Web of Science (1945–2013), the Cochrane Library Database (Issue 12, 2013), PubMed (1966–2013), EMBASE (1980–2013), CINAHL (1982–2013) and the Chinese Biomedical Database (CBM) (1982–2013) were searched for relevant articles without language restrictions. Meta-analysis was conducted using the STATA 12.0 software. Crude odds ratios (OR) with their corresponding 95 % confidence interval (CI) were calculated. Eight case–control studies with a total of 4,655 patients with cerebral infarction and 15,920 healthy control subjects were included in our meta-analysis. Five common polymorphisms in the LDLR gene were evaluated, including rs11669576 A > T, rs1433099 C > T, rs5925 C > T, rs688 C > T, rs1122608 T > G in the LDLR gene. The results of this meta-analysis revealed that cerebral infarction patients had a higher frequency of LDLR genetic polymorphisms than that of healthy controls (allele model: OR 1.17, 95 % CI 1.05–1.30, P = 0.004; dominant model: OR 1.18, 95 % CI 1.05–1.33, P = 0.007; homozygous model: OR 1.50, 95 % CI 1.03–2.16, P = 0.032; respectively), especially for the rs11669576 A > T, rs1433099 C > T and rs5925 C > T polymorphisms. Among different ethnic subgroups, the results demonstrated positive correlations between LDLR genetic polymorphisms and an increased risk of cerebral infarction among both Asians and Caucasians under the allele and dominant models (all P < 0.05). Our findings indicate that LDLR genetic polymorphisms may be strongly involved in the pathogenesis of cerebral infarction, especially the rs11669576 A > T, rs1433099 C > T, rs5925 C > T polymorphisms.     

Informed Conditioning on Clinical Covariates Increases Power in Case-Control Association Studies

Genetic case-control association studies often include data on clinical covariates, such as body mass index (BMI), smoking status, or age, that may modify the underlying genetic risk of case or control samples. For example, in type 2 diabetes, odds ratios for established variants estimated from low–BMI cases are larger than those estimated from high–BMI cases. An unanswered question is how to use this information to maximize statistical power in case-control studies that ascertain individuals on the basis of phenotype (case-control ascertainment) or phenotype and clinical covariates (case-control-covariate ascertainment). While current approaches improve power in studies with random ascertainment, they often lose power under case-control ascertainment and fail to capture available power increases under case-control-covariate ascertainment. We show that an informed conditioning approach, based on the liability threshold model with parameters informed by external epidemiological information, fully accounts for disease prevalence and non-random ascertainment of phenotype as well as covariates and provides a substantial increase in power while maintaining a properly controlled false-positive rate. Our method outperforms standard case-control association tests with or without covariates, tests of gene x covariate interaction, and previously proposed tests for dealing with covariates in ascertained data, with especially large improvements in the case of case-control-covariate ascertainment. We investigate empirical case-control studies of type 2 diabetes, prostate cancer, lung cancer, breast cancer, rheumatoid arthritis, age-related macular degeneration, and end-stage kidney disease over a total of 89,726 samples. In these datasets, informed conditioning outperforms logistic regression for 115 of the 157 known associated variants investigated (P-value = 1×10⁻⁹). The improvement varied across diseases with a 16% median increase in χ² test statistics and a commensurate increase in power. This suggests that applying our method to existing and future association studies of these diseases may identify novel disease loci.     

Regression Calibration in Semiparametric Accelerated Failure Time Models

Summary In large cohort studies, it often happens that some covariates are expensive to measure and hence only measured on a validation set. On the other hand, relatively cheap but error‐prone measurements of the covariates are available for all subjects. Regression calibration (RC) estimation method ( Prentice, 1982 , Biometrika 69 , 331–342) is a popular method for analyzing such data and has been applied to the Cox model by Wang et al. (1997, Biometrics 53 , 131–145) under normal measurement error and rare disease assumptions. In this article, we consider the RC estimation method for the semiparametric accelerated failure time model with covariates subject to measurement error. Asymptotic properties of the proposed method are investigated under a two‐phase sampling scheme for validation data that are selected via stratified random sampling, resulting in neither independent nor identically distributed observations. We show that the estimates converge to some well‐defined parameters. In particular, unbiased estimation is feasible under additive normal measurement error models for normal covariates and under Berkson error models. The proposed method performs well in finite‐sample simulation studies. We also apply the proposed method to a depression mortality study.     

Comparison of racial differences in childhood cancer risk in case-control studies and population-based cancer registries

Introduction: Although selection bias in case-control studies has been studied extensively, little is known about selection of cases and controls among various ethnic groups. This study compares racial differences in childhood cancer rates as estimated by case-control studies with various design features. It also compares estimates of racial distribution among cases as reported by case-control studies to those observed for an ideal case series with complete ascertainment of cases for these studies or in population-based cancer registries in corresponding geographic regions and calendar periods. Methods: Peer-reviewed publications on childhood leukemia and brain tumors from North America, published between 1980 and 2007, were reviewed. Incidence data by race/ethnicity were compiled from research publications, federal cancer statistics, and cancer registries. Meta-analysis was conducted to assess racial/ethnic differences by study characteristics. Racial distributions of cases from published case-control studies were compared to those of a presumably noncensored case distribution (i.e. include both participating and non-participating cases in a case-control study) or cases recorded by cancer registries. Results: In interview-based case-control studies of childhood cancer, the proportion of Whites compared to non-Whites tended to be higher among controls than among cases; however, the opposite was true for record-based case-control studies. Additionally, the proportion of Whites tended to be higher among the participating cases in the published case-control studies compared to the proportion of Whites among the non-participating cases or in cancer registries. Conclusions: Investigators need to consider differential participation by racial group as a potential source of bias in the interpretation of case-control study results.     

Effectiveness of two sampling protocols to survey orchid bees (Hymenoptera: Apidae) in the Neotropics

We compared the performances of two different methodologies for collecting orchid bees (Hymenoptera: Apidae): active sampling (at chemical baits with hand nets) versus passive sampling (with baited traps). Data were retrieved from previous field studies in which both methodologies were employed simultaneously. We evaluated the effectiveness of each method in terms of the number of species collected and the composition of the bee species through paired t test and Permutational Multivariate Analysis of Variance, respectively. Active sampling collected significantly more bee species than passive sampling in all but one study. The composition of bee species collected by each of the two methods was also significantly different. Larger bees belonging to Eulaema and Eufriesea species tended to be more frequent in traps. In contrast, all Euglossa species presented comparatively higher relative frequencies with active sampling. We challenge previous assertions that both methodologies are equally effective and argue that insisting on the sole use of baited traps in orchid-bee inventories may result in retrieving strongly biased raw data which will eventually lead to incorrect conclusions about the actual composition of orchid-bee assemblages. Since orchid bees have been more and more used as key organisms in studies on environmental impact in the Neotropics, particularly in Brazil, and the passive sampling methodology has been adopted as the “official” one by the Brazilian environmental agency (IBAMA), we suggest a re-evaluation of this practice and the adoption of the more expensive, but more reliable, methodology of using well-trained field researchers to actively collect data on these important Neotropical pollinators.     

Polymorphisms in the GNB3 and ADD1 genes and blood pressure in a Chinese population

A large proportion of the phenotypic variation in blood pressure (BP) appears to be inherited as a polygenic trait. This study examined the association between 12 single nucleotide polymorphisms (SNPs) in the guanine nucleotide binding protein beta polypeptide 3 (GNB3) and adducin 1 alpha (ADD1) genes and systolic (SBP), diastolic (DBP), and mean arterial (MAP) BP. A total of 3,142 individuals from 636 families were recruited from rural north China, and 2,746 met the eligibility criteria for analysis. BP measurements were obtained using a random-zero sphygmomanometer. Genetic variants were determined using SNPlex assays on an automated DNA Sequencer. A mixed linear model was used to estimate the association between each SNP and BP level. After Bonferroni correction, marker rs4963516 of the GNB3 gene remained significantly associated with DBP (corrected P values = 0.006, 0.007 and 0.002 for co-dominant, additive, and recessive models, respectively) and MAP (corrected P values = 0.02, 0.049, and 0.005, respectively). Compared to carriers of the major A allele, CC homozygotes had higher mean DBP (75.81 ± 0.62 vs. 73.46 ± 0.25 mmHg, P = 0.0002) and MAP (91.87 ± 0.68 vs. 89.42 ± 0.28 mmHg, P = 0.0004) after adjusting for covariates of age, gender, BMI, study site, and room temperature during BP measurement. In summary, these data support a role for the GNB3 gene in BP regulation in the Chinese population. Future studies aimed at replicating these novel findings are warranted.     

Risk of Sudden Sensorineural Hearing Loss in Patients with Common Preexisting Sensorineural Hearing Impairment: A Population-Based Study in Taiwan

Objective

The role of preexisting sensorineural hearing impairment on the risk for sudden sensorineural hearing loss (SSHL) is still unclear. In this study, we aimed to assess the risk of SSHL in patients with common preexisting sensorineural hearing impairment using population-based data.

Methods

A population-based case-control study design was used to analyze claims data between January 2001 and December 2011 obtained from the Taiwan National Health Insurance Research Database. The cases consisted of 514 patients with SSHL and the controls were frequency matched to 2,570 cases by sex, 10-year age group, and year of index date. Common sensorineural hearing impairments were retrospectively assessed in the cases and controls. Associations between sensorineural hearing impairment and risk of SSHL were evaluated using unconditional univariate and multivariate logistic regression analyses.

Results

The mean age for the 3,084 study subjects was 53.1 years (standard deviation, S.D. = 15.6). Of the 514 cases, 49 (9.5%) had sensorineural hearing impairment while only 44 (1.7%) of the 2,570 controls had the same condition. Univariate logistic regression analyses indicated that preexisting sensorineural hearing impairment was significantly associated with SSHL (odds ratio, OR = 6.05, p < 0.001). Other comorbidities including hypertension, diabetes mellitus, and hyperlipidemia also showed significant associations with SSHL. Similar results were obtained when the association between SSHL and sensorineural hearing impairment was adjusted with either all the covariates (adjusted OR = 6.22, p < 0.001) or with only those selected using a backward elimination procedure (adjusted OR = 6.20, p < 0.001).

Conclusions

Results from this population-based case-control study revealed that common sensorineural hearing impairment might be a novel risk factor for SSHL.     

Association between the rs6950982 polymorphism near the SERPINE1 gene and blood pressure and lipid parameters in a high-cardiovascular-risk population: interaction with Mediterranean diet

The SERPINE1 (serpin peptidase inhibitor, clade E, member 1) gene, better known by its previous symbol PAI-1 (plasminogen activator inhibitor 1), has been associated with cardiovascular phenotypes with differing results. Our aim was to examine the association between the rs6950982 (G > A) near the SERPINE1 gene, blood pressure (BP) and plasma lipid concentrations as well as the modulation of the polymorphism effects by adherence to Mediterranean diet (AMD). We studied 945 high-cardiovascular-risk subjects. Biochemical, clinical, dietary and genetic data (rs6950982) were obtained. We also determined the common rs1799768 (4G/5G), for checking independent effects. AMD was measured by a validated questionnaire, and four groups were considered. rs6950982 (A > G) and rs1799768 (4G/5G) were only in moderate–low linkage disequilibrium (D′ = 0.719; r ² = 0.167). The most significant associations we obtained were with rs6950982 (A > G). In males, the G allele was nominally associated with higher diastolic BP (AA: 81.5 ± 10.9, AG: 82.1 ± 11.4, GG: 85.7 ± 10.5 mmHg; P _additive = 0.030) and systolic BP (AA + AG: 141.4 ± 6.9 mmHg vs. GG: 149.8 ± 8.0 mmHg; P _recessive = 0.036). In the whole population, the rs6950982 was also associated with plasma lipids. Subject with the G allele presented higher total cholesterol (P _additive = 0.016, P _recessive = 0.011), low-density lipoprotein cholesterol (P _additive = 0.032, P _recessive = 0.031) and triglycerides (P _additive = 0.040, P _recessive = 0.029). AMD modulated the effect of rs6950982 on triglyceride concentrations (P for interaction = 0.036). Greater AMD reduced the higher triglyceride concentrations in GG subjects. No significant interactions were found for the other parameters. The rs6950982 was associated with higher BP in men and higher triglycerides in the whole population, this association being modulated by AMD.     

Meta-analysis of longitudinal studies: Serum vitamin D and prostate cancer risk

Aim: To review and summarize evidence from longitudinal studies on the association between serum 25-hydroxyvitamin D (25(OH)D) and the risk of prostate cancer (PC). Methods: Relevant prospective cohort studies and nested case-control studies published until July 2009 were identified by systematically searching Ovid Medline, EMBASE, and ISI Web of Knowledge databases and by cross-referencing. The following data were extracted in a standardized manner from eligible studies: first author, publication year, country, study design, characteristics of the study population, duration of follow-up, PC incidence/PC mortality according to serum vitamin D status and the respective risk ratios, and covariates adjusted for in the analysis. Due to the heterogeneity of studies in categorizing serum vitamin D levels, all results were recalculated for an increase in serum 25(OH)D by 10 ng/ml. Summary odds ratios (ORs) were calculated using meta-analysis methods. Results: Overall, eleven original articles were included, ten of which reported on the association between serum vitamin D levels and PC incidence and one article reported on the association with PC mortality. Meta-analysis of studies on PC incidence resulted in a summary OR (95% confidence interval, CI) of 1.03 (0.96–1.11) associated with an increase of 25(OH)D by 10 ng/ml (P = 0.362). No indication for heterogeneity and publication bias was found. Conclusions: According to available evidence from longitudinal studies, serum 25(OH)D is not associated with PC incidence.     

Evaluating Efficacy of Landsat-Derived Environmental Covariates for Predicting Malaria Distribution in Rural Villages of Vhembe District,South Africa

Malaria in South Africa is still a problem despite existing efforts to eradicate the disease. In the Vhembe District Municipality, malaria prevalence is still high, with a mean incidence rate of 328.2 per 100,0000 persons/year. This study aimed at evaluating environmental covariates, such as vegetation moisture and vegetation greenness, associated with malaria vector distribution for better predictability towards rapid and efficient disease management and control. The 2005 malaria incidence data combined with Landsat 5 ETM were used in this study. A total of nine remotely sensed covariates were derived, while pseudo-absences in the ratio of 1:2 (presence/absence) were generated at buffer distances of 0.5–20 km from known presence locations. A stepwise logistic regression model was applied to analyse the spatial distribution of malaria in the area. A buffer distance of 10 km yielded the highest classification accuracy of 82% at a threshold of 0.9. This model was significant (ρ < 0.05) and yielded a deviance (D²) of 36%. The significantly positive relationship (ρ < 0.05) between the soil-adjusted vegetation index and malaria distribution at all buffer distances suggests that malaria vector (Anopheles arabiensis) prefer productive and greener vegetation. The significant negative relationship between water/moisture index (a₁ index) and malaria distribution in buffer distances of 0.5, 10, and 20 km suggest that malaria distribution increases with a decrease in shortwave reflectance signal. The study has shown that suitable habitats of malaria vectors are generally found within a radius of 10 km in semi-arid environments, and this insight can be useful to aid efforts aimed at putting in place evidence-based preventative measures against malaria infections. Furthermore, this result is important in understanding malaria dynamics under the current climate and environmental changes. The study has also demonstrated the use of Landsat data and the ability to extract environmental conditions which favour the distribution of malaria vector (An. arabiensis) such as the canopy moisture content in vegetation, which serves as a surrogate for rainfall.