Notes Washed Up in a Bottle

Don. W. MacCorquodale M.D. M.S.P.H.

Periodic Notes From the Field on Philosophy and Science.

ALCOHOL AND BREAST CANCER RISK

          A number of epidemiologic studies have reported an increase in the risk of developing breast cancer associated with the consumption of alcohol.  Many of them demonstrated an increase in risk of about 30% - 50% with moderate alcohol consumption, one to two drinks per day.  (Terry M B et al. Lifetime alcohol intake and breast cancer risk.  Annals of Epidemiology 2006; 16:  230-240).  Questions remain about the effect of timing of alcohol intake and effect modification by other risk factors including body size, menopausal activity, and hormone therapy (HRT).
          The authors conducted a case-control study using data from the Long Island Breast Cancer Study Project.  Cases were women diagnosed with breast cancer between August 1, 1996 and July 31, 1997.  Controls were selected by random digit dialing and from the Health Care Financing Administration lists.
          In person interviews were conducted to assess subjects’ intake of alcohol intake by type, quantity, and frequency in each decade of life.  The authors used standard conversions applied by others of 13.2, 11.6, and 14.1 grams of alcohol for one 12 oz. bottle of beer, a 4 oz. glass of wine, and one 1.5 ounce serving of hard liquor.  Data were also collected regarding reproductive history, hormone use, menopausal status, BMI, cigarette smoking, family history of breast cancer, and demographic information.
          Unconditional logistic regression was employed to estimate odds ratios and 95% confidence intervals adjusting for potential confounding variables.   The associations between average lifetime alcohol intake and breast cancer risk are shown in the following table:

Lifetime alcohol             Cases          Controls
Intake                                      n = 1,508    n = 1,556    OR (95% CI)*

*adjusted for age at diagnosis, race, education, and BMI
+ reference group

          Current alcohol intake was not associated with increased risk of breast cancer.  “Analyses by type of alcohol (beer, wine, liquor) were not materially different and did not alter the conclusions for lifetime and current alcohol intake.”
          The association of lifetime intake of one or two drinks per day was limited to BMI < 25 (OR = 2.13; 95% CI:  1.29-3.54).  There was no association between heavier alcohol use and breast cancer risk.  Among postmenopausal women the association was higher among non HRT users (OR = 1.96; 95% CI:  1.10-3.48).  Lighter or heavier consumption was not associated with breast cancer risk.
          COMMENT:  Does lifetime, moderate consumption of alcohol increase a woman’s risk of breast cancer?  I wonder.  The lack of a dose-response effect makes me a bit uneasy about concluding that it does, despite the fact that there has been impressive consistency among many studies of the question.  The findings from this study imply that while moderate consumption of alcohol increases the risk of breast cancer, they also suggest that the ladies, who have knocked ‘em back with less restraint throughout their lives had done so with impunity.
          I found it impressive however that the women who consumed alcohol moderately were at increased risk of breast cancer among 1) those women with a BMI of less than 25, 2) postmenopausal women, 3) postmenopausal women, and 4) all women in the sample.  That adds credibility to the concept that moderate alcohol consumption is indeed causally associated with increased risk of breast cancer.  

MATERNAL DEPRESSION AND CHILD PSYCHOPATHOLOGY

          Parental depression is one of the most consistently noted risk factors for childhood anxiety, depression and disruptive behavior disorders.  A 2- to 3-fold increase has been reported among children of depressed parents as compared to controls (Weissmann M M et al.  Remissions in maternal depressions and child psychopathology.  JAMA 2006; 295:  1289-1298).  The authors examined “the symptomatic and behavioral functioning of children” evaluated 3 months after their mothers began treatment for depression.  They hypothesized that improvement in maternal depression would be associate with reductions in “psychophysiological symptoms and disorders in their children.” 
          The sample consisted of 151 mothers who were enrolled in STAR*D, a multi-site US study of different treatments for nonpsychotic major depressive disorder.  Participants had not been diagnosed with bipolar disorder, schizophrenia or schizoaffective disorders.  Participating women were screened to ascertain that they had children 7 to 17 years of age living with their mothers at last 50% of the time (in the case of divorced parents).
          Children’s psychiatric disorders at baseline and at the 3-month evaluation were established by direct interviews of the mothers and children.  Two outcome measures were examined:  change in overall rates of children’s’ diagnoses from baseline to 3 months and changes in the Child Behavior Checklist, CBCL.  Analyses were adjusted for age and sex of the child, severity of the mothers’ baseline symptoms, annual household income, and other potential confounders.
          One hundred fourteen (114) of the mother-child pairs remained in the study at the time of their child’s three months assessment.  The overall remission rate was 47%, 54/114.
          Mothers who remained depressed were 1) financially poorer, 2) more often on welfare, and 3) less likely to hold a college degree.  (No surprises here.  DWMac). 
          There was an overall 11% decrease in rates of diagnoses in children whose mothers had remissions of their depressions as opposed to an 8% increase in rates of diagnoses in children of mothers with continuing depression.  Thirty-seven, 37, children had psychiatric diagnoses at baseline.  Of those whose mothers’ depression remitted, one third (4/12) of the children’s diagnoses remitted.  Only 12% (3/25) of the children of women whose depression remained lost their diagnoses.   There difference was not statistically significant (p = .21).  (NB.  Note small sample sizes.  DWMac)
          Sixty-eight, 68, children had a psychiatric diagnosis at baseline.  Of these children, all remained free of psychiatric diagnoses if maternal depression remitted.  Of the children whose mothers remained depressed, 17% (8/46) had onset of diagnoses over this period (p = .05).
          Children whose mothers’ depression remitted showed a decrease in depressive behavior (18% to 9%), and disruptive behavior (18% to 12%).  There was no change in anxiety disorders.
          Among children of mothers who did not remit, there was an increase in the rates of depression (7% to 11%), anxiety (17o 25%), and disruptive behavior (20% to 24%).  Small sample size precluded statistical analysis of each disorder. 
          The authors stated that their findings suggest that remission of maternal depression over 30 months is statistically significantly associated with reduction in children’s current symptoms and diagnoses after controlling for child’s age, SES (annual household income), and severity of the mother’s depression at baseline.  They added that they could not demonstrate causality because their design was not experimental.
          COMMENT:  It would be helpful to see these findings replicated with a considerably larger sample.  I don’t think we can expect a randomized trial for ethical reasons.

LOW DENSITY LIPOPROTEIN CHOLESTEROL, LDL-C, AND CORONARY HEART DISEASE, CHD

          Causality regarding LDL-C levels and coronary heart disease is well established.  Differences in LDL-C levels are associated with 10-fold differences in CHD mortality (Brown M S and J L Goldstein.  Biomedicine.  Lowering LDL – Not only how low but how long?  Science 2006:  34:  1721-1723). 
          Heterozygous familial hypercholesterolemia with its mutations in the LDL receptor gene, doubles LDL-C levels throughout life, and it increases the risk of early heart attack by more than 10-fold.  It has been disappointing that treatment with cholesterol-lowering statins for 5 years reduces the incidence of heart attack just 40%.
          J. C. Cohen and his colleagues studied middle-aged Americans with life-long low LDL levels due to mutations, which stopped the functioning of the gene encoding PCSK9, an enzyme in the serine protease group.  In a small group of individuals with severe mutations, LDL-C was reduced by 38  mg/dl, and the prevalence of CHD declined by an astonishing 88%.  In a large number of subjects with less severe mutations, LDL-C levels were only reduced by 21 mg/dl, yet CHD incidence declined by 47%.
          A study of 1800 African Americans revealed that 2% carried one of two loss-of-function mutations in PCSK9.  Those subjects with either mutation had LDL levels 40% lower on average than those without them.  These mutations were rate in Caucasians, 0.1%.
          Analysis of data from a prospective study of slightly more than 15,700 Caucasians and African Americans from four U. S. communities revealed that among about 3,200 African Americans without a PCSK0 mutation, LDL-C levels average 128 mg/dl.  Slightly more than 300 subjects, 10%, developed CHD.  Among 85 African Americans with a PCSK9 mutation, LDL-C levels were reduced to 100 mg/dl.  Only one of these 85 individuals developed CHD, an 80% reduction.
          Among Caucasians in the same study, 300 individuals, 3%, had a mutation that lowered LDL-C levels by only 21 mg/dl.  Nonetheless, CHD incidence was reduced by  57%.
          Why does lowering LDL-C concentrations by 50 mg/dl by a PCSK9 mutation reduced CHD incidence by 80%, whereas a 40 mg/dl lowering with a statin reduces CHD prevalence by only 13%?  The most likely answer is duration.  Atherosclerosis is a chronic disease that begins in teenage years.  People with a genetic mutation that lowers LDL levels have had low LDL levels throughout their lives.  People on statin trials have had the LDL levels lowered for just 5 years.
          The lesson of PCSK9 is obvious.  If an 88% reduction in the incidence of CHD is to be obtained, LDL levels must be lowered before atherosclerosis becomes advanced.
          The physiological means to lower LDL levels is through a stringent diet, one low in cholesterol and saturated fat.  If diet fails, drugs including statins and cholesterol-adsorption inhibitors can be used.  Early intervention may well put an end to the epidemic of coronary heart disease that has ravaged western populations in the 20th century.