David F. Andrews

University of Toronto, Toronto, Ontario, Canada

Are you David F. Andrews?

Claim your profile

Publications (26)165.64 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Dysregulation of iron homeostasis is implicated in Alzheimer's disease (AD). In this pilot study, common variants of the apolipoprotein E (APOE) and HFE genes resulting in the iron overload disorder of hereditary hemochromatosis (C282Y, H63D and S65C) were evaluated as factors in sporadic AD in an Ontario sample in which folic acid fortification has been mandatory since 1998. Laboratory studies also were done to search for genetic effects on blood markers of iron status, red cell folates and serum B12. Participants included 58 healthy volunteers (25 males, 33 females) and 54 patients with probable AD (20 males, 34 females). Statistical analyses were interpreted at the 95% confidence level. Contingency table and odds ratio analyses supported the hypothesis that in females of the given age range, E4 significantly predisposed to AD in the presence but not absence of H63D. In males, E4 significantly predisposed to AD in the absence of H63D, and H63D in the absence of E4 appeared protective against AD. Among E4+ AD patients, H63D was associated with significant lowering of red cell folate concentration, possibly as the result of excessive oxidative stress. However, folate levels in the lowest population quartile did not affect the risk of AD. A model is presented to explain the experimental findings.
    Journal of Alzheimer's disease: JAD 05/2008; 14(1):69-84. · 3.61 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Dementia of the Alzheimer type (DAT) is common in older persons with Down syndrome (DS). There are three common alleles of the apolipoprotein E (ApoE) gene (Sigma 2, Sigma 3, and Sigma 4) resulting in three different isoforms (E2, E3, and E4) and six different genotypes (2,2; 2,3; 2,4; 3,3; 3,4; and 4,4). Sigma 4 is a risk factor for DAT whereas Sigma 2 appears prophylactic. As hepatitis B virus (HBV) infection and hypothyroidism also are common in DS, we evaluated associations between ApoE type, HBV status, and thyroid status in a sample of older persons with DS (n = 55; mean age, 44.3 +/- 10.8 years) using chi-squared analysis. Participants were classified as E2 (2,2 or 2,3), E3 (3,3), or E4 (3,4 or 4,4); positive for markers of HBV infection in the present or past (i.e., total HBcAb+ and/or HBsAg+ with or without infectivity, defined as HBV+) or negative for markers of HBV infection (defined as HBV-) and, currently receiving thyroid hormone supplement (defined as "hypothyroidism") or having normal thyroid function. The majority of the HBV+ were currently HBcAb+ and HBsAb+, but not HBsAg+. In females, there was an ApoE allele effect on thyroid status (P < or = 0.01), E2 being negatively (P < or = 0.01) and E4 being positively (P < or = 0.05) associated with "hypothyroidism". There was no evidence for an ApoE allele effect on thyroid status in males. There was no evidence for an ApoE allele effect on HBV status, or for an HBV status effect on thyroid status. As thyroid status can affect cognitive function, ApoE allele effects in DAT may, in part, be thyroid effects.
    American Journal of Medical Genetics Part A 07/2003; 120A(2):191-8. DOI:10.1002/ajmg.a.20099 · 2.05 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Familial aggregation of diseases potentially associated with metabolic syndrome (diabetes mellitus, hypertension, and cardiovascular diseases) was assessed in a colonoscopy-based case-control study of colorectal neoplasia in Toronto and Ottawa, Canada, in 1993-1996. Each familial disease was analyzed by logistic regression using generalized estimating equations. Case probands had incident adenomatous polyps (n = 172) or incident (n = 25) or prevalent (n = 132) colorectal cancer (CRC), while control probands (n = 282) had a negative colonoscopy and no history of CRC or polyps. Significant effect modification was evident in the data, with the strongest positive associations between familial diabetes and colorectal neoplasia among older probands with symptoms (parents: odds ratio (OR) = 2.4, 95% confidence interval (CI): 1.2, 4.8; siblings: OR = 5.8, 95% CI: 2.6, 13.3). Familial hypertension was also associated with colorectal neoplasia among probands with symptoms (OR = 1.7, 95% CI: 1.1, 2.6). In stratified analyses, familial diabetes, hypertension, and stroke were positively associated with adenomatous polyps in subgroups of probands who were older and/or had symptoms, while only familial diabetes was possibly associated with CRC. Associations in other proband groups may have been obscured by high cumulative incidence of parental CRC. Family studies are needed to understand the contribution of specific environmental and genetic factors in accounting for the disease aggregations.
    American Journal of Epidemiology 11/2002; 156(8):702-13. · 4.98 Impact Factor
  • Source
    American Journal of Medical Genetics 12/2000; 95(2):189. DOI:10.1002/1096-8628(20001113)95:23.0.CO;2-7 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Mutations in the class I-like major histocompatibility complex gene called HFE are associated with hereditary hemochromatosis (HHC), a disorder of excessive iron uptake. We screened DNA samples from patients with familial Alzheimer disease (FAD) (n = 26), adults with Down syndrome (DS) (n = 50), and older (n = 41) and younger (n = 52) healthy normal individuals, for two HHC point mutations-C282Y and H63D. Because the apolipoprotein E (ApoE) E4 allele is a risk factor for AD and possibly also for dementia of the AD type in DS, DNA samples were also ApoE genotyped. Chi-squared analyses were interpreted at the 0.05 level of significance without Bonferroni corrections. In the pooled healthy normal individuals, C282Y was negatively associated with ApoE E4, an effect also apparent in individuals with DS but not with FAD. Relative to older normals, ApoE E4 was overrepresented in both males and females with FAD, consistent with ApoE E4 being a risk factor for AD; HFE mutations were overrepresented in males and underrepresented in females with FAD. Strong gender effects on the distribution of HFE mutations were apparent in comparisons among ApoE E4 negative individuals in the FAD and healthy normal groups (P < 0.002). Our findings are consistent with the proposition that among ApoE E4 negative individuals HFE mutations are predisposing to FAD in males but are somewhat protective in females. Further, ApoE E4 effects in our FAD group are strongest in females lacking HFE mutations. Relative to younger normals there was a tendency for ApoE E4 and H63D to be overrepresented in males and underrepresented in females with DS. The possibility that HFE mutations are important new genetic risk factors for AD should be pursued further.
    American Journal of Medical Genetics 07/2000; 93(1):58-66. DOI:10.1002/1096-8628(20000703)93:1<58::AID-AJMG10>3.0.CO;2-L · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this study was to determine the risk of endoscopic/radiological recurrence of Crohn's disease postoperatively and the long-term outcome. A randomized placebo-controlled trial was performed to determine the effectiveness of mesalamine in preventing recurrent Crohn's disease postoperatively. Patients in the control group were examined endoscopically/radiologically before entry into and annually during the trial. Findings were classified as minimal or severe. There were 76 patients (49 men and 37 women; mean age, 37.1 +/- 13.2 years). Fifty (61.7%) had terminal ileal resections. Overall, 55 endoscopic/radiological recurrences were observed in 51 patients (67.1%). Expressed actuarially, the recurrence rate was 27.5% at 1 year (95% confidence interval [CI], 15.8%-37.6%), 60.8% at 2 years (95% CI, 46%-71.3%), and 77.3% at 3 years (95% CI, 62.7%-86.3%). Nineteen (37%) were symptomatic and 12 (24%) were initially asymptomatic but later became symptomatic (mean, 13.0 +/- 8.8 months), whereas 20 (39%) remained asymptomatic (mean, 16.9 +/- 17.4 months). Patients with severe endoscopic/radiological disease were significantly more likely to be or become symptomatic than those with minimal disease (23 of 32 vs. 8 of 19, respectively; P = 0.0437). This study suggests that postoperative endoscopic/radiological recurrences occur later than previously reported. Furthermore, many of these patients, especially with minimal disease, will remain asymptomatic.
    Gastroenterology 01/1998; 113(6):1823-7. DOI:10.1016/S0016-5085(97)70001-5 · 13.93 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Recurrence of Crohn's disease frequently occurs after surgery. A randomized controlled trial was performed to determine if mesalamine is effective in decreasing the risk of recurrent Crohn's disease after surgical resection is performed. One hundred sixty-three patients who underwent a surgical resection and had no evidence of residual disease were randomized to a treatment group (1.5 g mesalamine twice a day) or a placebo control group within 8 weeks of surgery. The follow-up period was a maximum of 72 months. The symptomatic recurrence rate (symptoms plus endoscopic and/or radiological confirmation of disease) in the treatment group was 31% (27 of 87) compared with 41% (31 of 76) in the control group (P = 0.031). The relative risk of developing recurrent disease was 0.628 (90% confidence interval, 0.40-0.97) for those in the treatment group (P = 0.039; one-tail test) using an intention-to-treat analysis and 0.532 (90% confidence interval, 0.32-0.87) using an efficacy analysis. The endoscopic and radiological rate of recurrence was also significantly decreased with relative risks of 0.654 (90% confidence interval, 0.47-0.91) in the effectiveness analysis and 0.635 (90% confidence interval, 0.44-0.91) in the efficacy analysis. There was only one serious side effect (pancreatitis) in subjects in the treatment group. Mesalamine (3.0 g/day) is effective in decreasing the risk of recurrence of Crohn's disease after surgical resection is performed.
    Gastroenterology 09/1995; 109(2):404-13. DOI:10.1016/0016-5085(95)90327-5 · 13.93 Impact Factor
  • James E. Stafford, David F. Andrews, Yong Wang
    [Show abstract] [Hide abstract]
    ABSTRACT: We describe a set of procedures that automate many algebraic calculations common in statistical asymptotic theory. The procedures are very general and serve to unify the study of likelihood and likelihood type functions. The procedures emulate techniques one would normally carry out by hand; this strategy is emphasised throughout the paper. The purpose of the software is to provide a practical alternative to difficult manual algebraic computations. The result is a method that is quick and free of clerical error.
    Statistics and Computing 01/1994; 4(4):235-245. DOI:10.1007/BF00156746 · 1.75 Impact Factor
    [Show abstract] [Hide abstract]
    ABSTRACT: A symbolic algorithm for the derivation of the asymptotic expansion of the maximum likelihood estimate was first introduced by Andrews and Stafford (1993). This algorithm is quite general and can be modified to give asymptotic expansions for the maximum likelihood estimate based on either the conditional, modified or adjusted profile likelihoods. Expansions are obtained quickly with no chance for human clerical error. As a result, these three adjustments can be easily studied and some new statistical results are obtained. Bartlett corrections of each adjusted likelihood, and asymptotic expansions for the first four moments of the maximum likelihood estimate based on the conditional and adjusted profile likelihoods, are derived.
    Biometrika 12/1993; 80(4):715-730. DOI:10.1093/biomet/80.4.715 · 1.51 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We previously observed low level mosaicism (2-4% normal cells) in phytohemagglutinin-stimulated peripheral blood lymphocytes (PBL) in 29% of a small group of elderly persons with Down syndrome (DS). An analysis of cytogenetic data on 154 trisomy 21 cases (age 1 day to 68 years) showed that the proportion of diploid cells in such cultures significantly increased (P < 0.005) with advancing age. Thus, the "occult" mosaicism in PBL of the elderly persons with DS is likely due to the accumulation of cells that have lost a chromosome 21. A consequence of chromosome 21 loss could be uniparental disomy of the 2n cells, a factor that might have significant biological consequences if some chromosome 21 genes are imprinted. Loss of a chromosome 21 from trisomic cells might result in tissue-specific mosaicism and "classical" mosaicism in different age groups. Chromosome 21 loss might also be relevant to the development of Alzheimer-type dementia in DS and in the general population.
    American Journal of Medical Genetics 03/1993; 45(5):584-8. DOI:10.1002/ajmg.1320450513 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Probable Alzheimer disease (AD) is described in 79-year-old male twins with monozygosity confirmed by DNA examination. The first twin to be affected began to show signs of intellectual deterioration at age 70. In the other, onset was at age 72. Four of their living sibs (current age range = 75-92) are also suspected to have AD. The possible roles of genetic and environmental factors in the development of AD in this sibship are discussed.
    American Journal of Medical Genetics 11/1992; 44(5):591-7. DOI:10.1002/ajmg.1320440512 · 3.23 Impact Factor
  • Source
    The Lancet 08/1991; 338(8762):324–326. DOI:10.1016/0140-6736(91)90472-2 · 39.21 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Although epidemiological and biochemical evidence suggests that aluminium may be associated with Alzheimer's disease (AD), there is no convincing proof of a causal link for aluminium in disease progression. We have completed a two year, single-blind study to investigate whether the progression of dementia could be slowed by the trivalent ion chelator, desferrioxamine. 48 patients with probable AD were randomly assigned to receive desferrioxamine (125 mg intramuscularly twice daily, 5 days per week, for 24 months), oral placebo (lecithin), or no treatment. No significant differences in baseline measures of intelligence, memory, or speech ability existed between groups. Activities of daily living were assessed and videorecorded at 6, 12, 18, and 24 month intervals. There were no differences in the rate of deterioration of patients receiving either placebo or no treatment. Desferrioxamine treatment led to significant reduction in the rate of decline of daily living skills as assessed by both group means (p = 0.03) and variances (p less than 0.04). The mean rate of decline was twice as rapid for the no-treatment group. Appetite (n = 4) and weight (n = 1) loss were the only reported side-effects. We conclude that sustained administration of desferrioxamine may slow the clinical progression of the dementia associated with AD.
    The Lancet 07/1991; 337(8753):1304-8. DOI:10.1016/0140-6736(91)92978-B · 39.21 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have identified 2 sisters with probable dementia of the Alzheimer type who have an unusual 22-derived marker chromosome with a greatly elongated short arm containing 2 well-separated nucleolus organizer regions. A marker chromosome similar in appearance is uncommon in the general population. Eleven of 24 of their biological relatives were also found to have the marker. The known pedigree of this family encompasses 6 generations in 2 of which there is evidence of 10 cases of dementia of the Alzheimer type. The average age-at-onset of dementia is 65.8 +/- 5.5 years; the average age-at-death among those apparently affected is 74.9 +/- 8.3 years. A new model for the estimation of risk was applied to the family data. Persons in this family with the marker were found to be 4 times more likely to develop dementia than those without the marker, the 95% confidence interval for this risk being 1-50. The probability that the association of dementia and the marker is due to chance alone is .05 (1 in 20).
    American Journal of Medical Genetics 06/1991; 39(3):307-13. DOI:10.1002/ajmg.1320390312 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Serum tests of thyroid function were compared in Down syndrome (DS) patients with and without manifestations of Alzheimer disease (AD). Relative to control individuals, DS patients had, overall, lower mean total T4 (P = 0.070) and T3f (P = 0.015), higher T3U (P = 0.013) and TSH (P = 0.020), no difference in free T4, and higher thyroid antithyroglobulin (ATA) (P = 0.033) and antimicrosomal autoantibody (AMA) titres (P = 0.0097). Similar trends were apparent in DS males and females, and in DS patients off all drugs. In an analysis of case/control pairs with corrections for age and sex, DS patients with AD manifestations (n = 9) had significantly lower T3 (P = 0.029) and higher AMA (P = 0.043) than paired control individuals, whereas DS patients without AD manifestations (n = 20) had significantly lower T3 (P = 0.013) but higher ATA (P = 0.0065). T3 was significantly lower in the DS patients with AD manifestations than in the unaffected (P = 0.0013). These data suggest that autoimmune thyroiditis associated with a mild "subclinical" form of hypothyroidism is common in adult DS patients and more pronounced in patients with AD manifestations than in those without. This "subclinical" hypothyroidism may contribute to cognitive deficits in ageing DS patients.
    American Journal of Medical Genetics 06/1990; 36(2):148-54. DOI:10.1002/ajmg.1320360205 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The activities of red blood cell enzymes that scavenge the superoxide radical and hydrogen peroxide were measured in severely to profoundly retarded adult Down syndrome (DS) patients with and without manifestations of Alzheimer disease (AD), and control individuals matched for sex, age, and time of blood sampling. Cu,Zn superoxide dismutase (SOD-1) and glutathione peroxidase (GSHPx) activities were significantly elevated (1.39-fold and 1.24-fold, respectively) in DS individuals without AD. When an adjustment was made for the SOD gene dosage effect, DS patients with AD manifestations had significantly lower SOD levels than the matched control individuals. In contrast, DS patients with and without AD had a similar elevation in GSHPx (an adaptive phenomenon). The mean catalase (CAT) activity was no different in DS and control individuals; however, in a paired regression analysis, DS patients without AD had marginally lower CAT activity than control individuals, whereas DS patients with AD had slightly but not significantly higher CAT activity. Thus, AD manifestations in this DS population are associated with changes in the red cell oxygen scavenging processes.
    American Journal of Medical Genetics 04/1990; 35(4):459 - 467. DOI:10.1002/ajmg.1320350403 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have begun a systematic study of human papillomavirus (HPV) infection in colposcopically and/or morphologically normal epithelium of the uterine cervix. Paired biopsies were taken from the lesions (cervical intraepithelial neoplasia [CIN], condyloma, invasive carcinoma) and from the normal-appearing adjacent epithelium 3 to 5 mm from the edge of the lesion. Myometrium or ectocervical epithelium from patients who had undergone hysterectomy for reasons other than genital dysplasia or malignancy served as controls. One biopsy was examined histologically. DNA from the second biopsy was digested with Pst I, and the presence or absence of HPV was determined by Southern blotting using HPV-16 DNA as a probe. HPV was not detected in any of the 12 control samples. Of 30 patients with CIN and/or condyloma, five of 18 who were HPV-positive had either HPV-16 (three cases) or virus resembling HPV-31 (two cases) in the lesion and adjacent epithelium. Of seven patients with invasive carcinoma, four had HPV in the lesion and adjacent epithelium; two of these four patients had typical HPV-16. Such infection of apparently normal epithelium has major implications for our understanding of the pathogenesis, treatment, and follow-up of patients with cervical neoplasia.
    Human Pathlogy 05/1989; 20(4):316-9. DOI:10.1016/0046-8177(89)90039-7 · 2.81 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In carrier detection studies, females-at-risk are usually tested several times if the results are ambiguous, whereas subjects in the control and obligate carrier reference groups may not be tested as often. The question is how to incorporate the multiple measurements most effectively with information in the family pedigree into combined carrier risks. Sets of measurements on individuals are not independent, but are related with the correlation coefficient 0 less than rho less than 1. We have developed a procedure for incorporating repeated measurements on individuals and their a priori chance of having the disease into logistic models. This procedure utilizes the set of measurements and an estimate of rho. We describe application of this procedure to carrier detection in Duchenne muscular dystrophy (DMD) using serum creatine kinase (CK) measurements as the biochemical indicator of carrier status. Estimates of rho for controls and obligate DMD carriers did not differ significantly from 0.5. Repeated testing with use of rho = 0.5 significantly decreased the median logistic carrier probability for controls and increased it for carriers. In some cases four to six rather than the three CK tests conventionally used in genetic counseling were necessary to obtain a stable logistic carrier probability for a subject.
    American Journal of Medical Genetics 04/1987; 26(4):851-61. DOI:10.1002/ajmg.1320260412 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The program DUCHEN calculates the probability that a woman is a carrier of an X-linked, lethal recessive disease on the basis of information in the woman's family and any available biochemical data. It is easily used by persons without computer knowledge or experience. The present version can accommodate families consisting of up to 100 people in seven generations. Risks may be estimated on the basis of pedigree information only, or with the inclusion of one or more types of biochemical test results. Biochemical data are incorporated with pedigree information into final risks using the powerful statistical technique of logistic discrimination, a procedure particularly suited for the separation of non-normal populations on the basis of overlapping quantitative characteristics. Mutation rates are specified separately for males and females. DUCHEN is available in FORTRAN 77, IBM BASIC, and Applesoft BASIC, and may be used on a variety of mainframe or microcomputers. The model was used to calculate risks for 375 girls and women in 46 families with Duchenne muscular dystrophy (DMD); serum creatine kinase tests had been carried out on 167 of these subjects who were of reproductive age. Carrier probabilities equal to or lower than the population risk (0.0004) were obtained for 21% of the aunts and 43% of the cousins of affected boys from families with an isolated case of DMD and for 14% of the cousins of affected boys from families with a known DMD history. DUCHEN should assist counsellors in determining which members of large families should be further examined using either standard biochemical carrier detection methods or DNA marker studies.(ABSTRACT TRUNCATED AT 250 WORDS)
    American Journal of Medical Genetics 10/1986; 25(2):211-8. DOI:10.1002/ajmg.1320250203 · 3.23 Impact Factor
  • D. F. Andrews, R. Brant, M. E. Percy
    [Show abstract] [Hide abstract]
    ABSTRACT: A common problem in medical statistics is the discrimination between two groups on the basis of diagnostic information. Information on patient characteristics is used to classify individuals into one of two groups: diseased or disease-free. This classification is often with respect to a particular disease. This discrimination has two probabilistic components: (1) the discrimination is not without error, and (2) in many cases the a priori chance of disease can be estimated. Logistic models (Cox 1970; Anderson 1972) provide methods for incorporating both of these components. The a posteriori probability of disease may be estimated for a patient on the basis of both current measurement of patient characteristics and prior information. The parameters of the logistic model may be estimated on the basis of a calibration trial. In practice, not one but several sets of measurements of one characteristic of the patient may be made on a questionable case. These measurements typically are correlated; they are far from independent. How should these correlated measurements be used? This paper presents a method for incorporating several sets of measurements in the classification of a case.En statistique médicale, un problème commun consiste à séparer deux groupes d'individus à partir de l'information recueillie lors d'un examen médical. Cette information permet de classifier les patients entre malades et bien portants à partir de leur état physique. La classification, qui s'effectue habituellement en fonction d'une maladie spécifique, comporte deux composantes probabilistes bien distinctes: (1) la discrimination est sujette à crrcur; (2) dans plusieurs cas, on peut évaluer à priori le risque de contracter la maladie. Les modèles logistiques (voir Cox 1970 et Anderson 1972) permettent d'incorporer simultanément ces deux composantes. On peut alors réviser la probabilité qu'un patient ait contracté la maladie à l'aide de données recentes sur son état physique. Les paramètres inhérents au modèle logistique peuvent quant à eux ětre estimés après un test clinique de calibration. En pratique, il peut arriver qu'en cas de doute, on ait non pas un, mais plusieurs ensembles de données sur l'état d'un patient. Ces données sont habituellement corrélees; elles sont loin d'ětre indépendantes. Comment ces données corrélées devraient-elles ětre utilisées? Cet article répond à la question en présentant une méthode qui permet de traiter un problème de classification au moyen de plusieurs ensembles de données.
    Canadian Journal of Statistics 08/1986; 14(3):263 - 266. DOI:10.2307/3314805 · 0.59 Impact Factor

Publication Stats

931 Citations
165.64 Total Impact Points


  • 1981–2003
    • University of Toronto
      • • Department of Physiology
      • • Department of Surgery
      • • Department of Obstetrics and Gynaecology
      Toronto, Ontario, Canada
  • 1982
    • SickKids
      Toronto, Ontario, Canada