Low serum cholesterol and the risk of cancer: an analysis of the published prospective studies

Sequence variation in proprotein convertase subtilisin/kexin type 9 serine protease gene, low LDL cholesterol, and cancer incidence

Some prospective epidemiologic studies have suggested that a low plasma cholesterol level may be associated with increased risk of cancer. Certain sequence variants in the proprotein convertase subtilisin/kexin type 9 serine protease gene (PCSK9) are associated with lifelong low total and LDL cholesterol. We therefore analyzed the association of PCSK9 variation with incidence of cancer between 1987 and 2000 in a prospective study (n=13,250). The frequency of the PCSK9 variants studied was 2.4% in blacks and 3.2% in whites. Neither was associated with increased cancer incidence: age- and sex-adjusted hazard ratios were 0.66 [95% confidence interval (95% CI), 0.31-1.39] in blacks and 0.77 (95% CI, 0.54-1.09) in whites. Low baseline total or LDL cholesterol levels in 1987 to 1989 were also not statistically significantly associated with incident cancer: multivariable-adjusted hazard ratios for the lowest compared with the highest quartiles of LDL cholesterol were 1.05 (95% CI, 0.78-1.40) in blacks and 1.16 (95% CI, 0.99-1.36) in whites. These data suggest that a lifelong low cholesterol concentration, as reflected by these PCSK9 variants, does not increase risk of cancer.

Lipids changes in liver cancer

Liver is one of the most important organs in energy metabolism. Most plasma apolipoproteins and endogenous lipids and lipoproteins are synthesized in the liver. It depends on the integrity of liver cellular function, which ensures homeostasis of lipid and lipoprotein metabolism. When liver cancer occurs, these processes are impaired and the plasma lipid and lipoprotein patterns may be changed. Liver cancer is the fifth common malignant tumor worldwide, and is closely related to the infections of hepatitis B virus (HBV) and hepatitis C virus (HCV). HBV and HCV infections are quite common in China and other Southeast Asian countries. In addition, liver cancer is often followed by a procession of chronic hepatitis or cirrhosis, so that hepatic function is damaged obviously on these bases, which may significantly influence lipid and lipoprotein metabolism in vivo. In this review we summarize the clinical significance of lipid and lipoprotein metabolism under liver cancer.

Statin use and the risk of 10 cancers

BACKGROUND

Statins affect the proliferation, survival, and migration of cancer cells, and it is thought that they may have chemopreventive properties in humans. The purpose of the present study was to evaluate the association between statin use and various types of cancer in our hospital-based case-control surveillance study.

METHODS

Data were collected from patients ages 40-79 years who were admitted to participating hospitals in 3 centers in Philadelphia, New York, and Baltimore from 1991 to 2005. Nurses administered questionnaires to obtain information on medication use and other factors. We compared patients who had any of 10 types of cancer (a total of 4913 patients) with controls admitted for noncancer diagnoses (3900 patients). The following cancers were examined individually: female breast (n = 1185), prostate (n = 1226), colorectal (n = 734), lung (n = 464), bladder (n = 240), leukemia (n = 254), pancreas (n = 220), kidney (n = 226), endometrial (n = 220), and non-Hodgkin lymphoma (n = 144). Logistic regression models were used to estimate odds ratios and 95% confidence intervals among regular statin users compared with never-users.

RESULTS

Odds ratios were compatible with 1.0 for all cancer types. For the 4 largest cancer sites (breast, prostate, colorectum, and lung), odds ratios did not vary significantly by duration of statin use.

CONCLUSIONS

Statins are among the most commonly used medications, and durations of use are increasing. The present data do not support either positive or negative associations between statin use and the occurrence of 10 cancer types. Cancer incidence should continue to be monitored among statin users.

Statin use and risk of colorectal cancer

BACKGROUND

Statins have anticancer activity in various cell types, including colon cancer cells. Although epidemiologic data on the relationship between statin use and the risk of colorectal cancer are limited, one case-control study reported a 50% reduction in colorectal cancer risk among statin users. We conducted a population-based case-control study to assess this association with respect to statin type, dose, and duration of use.

METHODS

Case patients with adenocarcinoma of the colon or rectum were ascertained from participating hospitals in Massachusetts and the Massachusetts Cancer Registry from January 1, 2001, through November 30, 2004. Age-, sex-, and precinct-matched control subjects were chosen from Massachusetts town lists. Information on statin use and other relevant data were obtained by telephone interview. We used multivariable conditional logistic regression models to estimate odds ratios (ORs). All tests for statistical significance were two-sided.

RESULTS

Among 1809 case patients and 1809 matched control subjects, regular use of statins for at least 3 months was not associated with the risk of colorectal cancer (OR = 0.92, 95% confidence interval [CI] = 0.78 to 1.09). There was no consistent trend across dose or duration of use (e.g., for > or = 10 years of use, OR = 0.86, 95% CI = 0.51 to 1.45). The risk of stage IV cancer was, however, statistically significantly lower among statin users than among nonusers (OR = 0.49, 95% CI = 0.26 to 0.91). There was no evidence of an interaction between statin use and nonsteroidal anti-inflammatory drug use.

CONCLUSIONS

Overall, use of statins did not appear to be associated with reduced risk of colorectal cancer. The reduced risk of stage IV cancer observed among statin users requires confirmation.

Genetic polymorphism of cholesterol 7alpha-hydroxylase (CYP7A1) and colorectal adenomas: Self Defense Forces Health Study

Bile acids have long been implicated in colorectal carcinogenesis, but epidemiological evidence is limited. Cholesterol 7alpha-hydroxylase (CYP7A1) is the rate-limiting enzyme producing bile acids from cholesterol. A recent case-control study showed a decreased risk of proximal colon cancer associated with the CC genotype of the CYP7A1 A-203C polymorphism. The present study examined the relationship between the CYP7A1 A-203C polymorphism and colorectal adenoma, which is a well-established precursor lesion of colorectal cancer. The study subjects comprised 446 cases of colorectal adenomas and 914 controls of normal total colonoscopy among men receiving a preretirement health examination at two hospitals of the Self Defense Forces (SDF). The CYP7A1 genotype was determined by the polymerase chain reaction-restriction fragment length polymorphism method. Statistical adjustment was made for age, hospital, rank in the SDF, smoking, alcohol use, body mass index, physical activity and parental history of colorectal cancer. The CYP7A1 polymorphism was not measurably related to the overall risk of colorectal adenomas. However, the CC genotype was associated with a decreased risk of proximal colon adenomas, but not of distal colon and rectal adenomas. Adjusted odds ratios of proximal colon adenomas (95% confidence intervals) for the AC and CC genotype versus AA genotype were 0.82 (0.54-1.24) and 0.56 (0.34-0.95), respectively. The findings add to evidence for the role of bile acids in colorectal carcinogenesis. The CC genotype of the CYP7A1 A-203C polymorphism probably renders lower activity of the enzyme synthesizing bile acids.

Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins

BACKGROUND

Results of previous randomised trials have shown that interventions that lower LDL cholesterol concentrations can significantly reduce the incidence of coronary heart disease (CHD) and other major vascular events in a wide range of individuals. But each separate trial has limited power to assess particular outcomes or particular categories of participant.

METHODS

A prospective meta-analysis of data from 90,056 individuals in 14 randomised trials of statins was done. Weighted estimates were obtained of effects on different clinical outcomes per 1.0 mmol/L reduction in LDL cholesterol.

FINDINGS

During a mean of 5 years, there were 8186 deaths, 14,348 individuals had major vascular events, and 5103 developed cancer. Mean LDL cholesterol differences at 1 year ranged from 0.35 mmol/L to 1.77 mmol/L (mean 1.09) in these trials. There was a 12% proportional reduction in all-cause mortality per mmol/L reduction in LDL cholesterol (rate ratio [RR] 0.88, 95% CI 0.84-0.91; p<0.0001). This reflected a 19% reduction in coronary mortality (0.81, 0.76-0.85; p<0.0001), and non-significant reductions in non-coronary vascular mortality (0.93, 0.83-1.03; p=0.2) and non-vascular mortality (0.95, 0.90-1.01; p=0.1). There were corresponding reductions in myocardial infarction or coronary death (0.77, 0.74-0.80; p<0.0001), in the need for coronary revascularisation (0.76, 0.73-0.80; p<0.0001), in fatal or non-fatal stroke (0.83, 0.78-0.88; p<0.0001), and, combining these, of 21% in any such major vascular event (0.79, 0.77-0.81; p<0.0001). The proportional reduction in major vascular events differed significantly (p<0.0001) according to the absolute reduction in LDL cholesterol achieved, but not otherwise. These benefits were significant within the first year, but were greater in subsequent years. Taking all years together, the overall reduction of about one fifth per mmol/L LDL cholesterol reduction translated into 48 (95% CI 39-57) fewer participants having major vascular events per 1000 among those with pre-existing CHD at baseline, compared with 25 (19-31) per 1000 among participants with no such history. There was no evidence that statins increased the incidence of cancer overall (1.00, 0.95-1.06; p=0.9) or at any particular site.

INTERPRETATION

Statin therapy can safely reduce the 5-year incidence of major coronary events, coronary revascularisation, and stroke by about one fifth per mmol/L reduction in LDL cholesterol, largely irrespective of the initial lipid profile or other presenting characteristics. The absolute benefit relates chiefly to an individual's absolute risk of such events and to the absolute reduction in LDL cholesterol achieved. These findings reinforce the need to consider prolonged statin treatment with substantial LDL cholesterol reductions in all patients at high risk of any type of major vascular event.

Genetic polymorphism in cytochrome P450 7A1 and risk of colorectal cancer: the Fukuoka Colorectal Cancer Study

Bile acids have long been implicated in the etiology of colorectal cancer, but epidemiologic evidence remains elusive. Cholesterol 7alpha-hydroxylase (CYP7A1) is the rate-limiting enzyme in the synthesis of bile acids from cholesterol in the liver, and thus may be an important determinant of bile acid production. We examined the association between the CYP7A1 A-203C polymorphism and colorectal cancer. The CYP7A1 A-203C polymorphism was determined by the PCR-RFLP method in 685 incident cases of colorectal cancer and 778 controls randomly selected from a community in the Fukuoka area, Japan. The CC genotype was slightly less frequent in the case group, and the adjusted odds ratio for the CC versus AA genotype was 0.88 (95% confidence interval, 0.65-1.20). In the analysis by subsite of the colorectum, a decreased risk associated with the CYP7A1 CC genotype was observed for proximal colon cancer, but not for either distal colon or rectal cancer. The adjusted odds ratios (95% confidence intervals) of proximal colon cancer for the CC genotype were 0.63 (0.36-1.10) compared with the AA genotype, and 0.59 (0.37-0.96) compared with the AA and AC genotypes combined. A decreased risk of proximal colon cancer in relation to the CC genotype of CYP7A1 A-203C, which probably renders less activity of the enzyme converting cholesterol to bile acids, is new evidence for the role of bile acids in colorectal carcinogenesis.

Long-term effectiveness and safety of pravastatin in 9014 patients with coronary heart disease and average cholesterol concentrations: the LIPID trial follow-up

BACKGROUND

The Long-term Intervention with Pravastatin in Ischaemic Disease (LIPID) study showed that pravastatin therapy over 6 years reduced mortality and cardiovascular events in patients with previous acute coronary syndromes and average cholesterol concentrations. We assessed the longer-term effects of initial treatment with pravastatin on further cardiovascular events and mortality over a total follow-up period of 8 years.

METHODS

In the main trial, 9014 patients with previous myocardial infarction or unstable angina and a baseline plasma cholesterol concentration of 4.0-7.0 mmol/L were randomly assigned pravastatin 40 mg daily or placebo and followed up for 6 years. Subsequently, all patients were offered open-label pravastatin for 2 more years. Major cardiovascular events and adverse events were compared according to initial treatment assignment.

FINDINGS

7680 (97% of those still alive) had 2 years of extended follow-up. 3766 (86%) of those assigned placebo and 3914 (88%) assigned pravastatin agreed to take open-label pravastatin. During this period, patients originally assigned pravastatin had almost identical cholesterol concentrations to those assigned placebo, but a lower risk of death from all causes (219 [5.6%] vs 255 [6.8%], p=0.029), coronary heart disease (CHD) death (108 [2.8%] vs 137 [3.6%], p=0.026), and CHD death or non-fatal myocardial infarction (176 [4.5%] vs 196 [5.2%], p=0.08). Over the total 8-year period, all-cause mortality was 888 (19.7%) in the group originally assigned placebo and 717 (15.9%) in the group originally assigned pravastatin, CHD mortality was 510 (11.3%) versus 395 (8.8%), myocardial infarction was 570 (12.7%) versus 435 (9.6%; each p < 0.0001), and stroke was 272 (6.0%) versus 224 (5.0%; p=0.015). Stronger evidence of separate treatment benefits than in the main trial was seen in important prespecified subgroups (women, patients aged > or = 70 years, and those with total cholesterol < 5.5 mmol/L). Pravastatin had no significant adverse effects.

INTERPRETATION

The evidence of sustained treatment benefits and safety of long-term pravastatin treatment reinforces the importance of long-term cholesterol-lowering treatment for almost all patients with previous CHD events.

How well tolerated are lipid-lowering drugs?

It has been clearly established that lipid-lowering treatments [such as 3-hydroxyl-3-methylglutamyl coenzyme A reductase inhibitors ('statins') or fibrates] can reduce cardiovascular events, and with one of the statins even total mortality, in high-risk populations. Intervention studies have not included the very old, but it is generally assumed that this patient group would benefit from these treatments to an extent similar to younger patients. Worries about the associations seen in observational studies between low cholesterol levels and cancer, cerebral haemorrhage or mood and behaviour change have been largely overcome by findings from the latest large drug intervention trials, which do not show any increase in these conditions with statin or fibrate treatments. The common adverse effects associated with these drugs are relatively mild and often transient in nature. Potentially more serious adverse effects, which are more clearly related to drug treatment and are probably dose-dependent, include elevations in hepatic transaminase levels and myopathy; however, these effects are uncommon and generally resolve rapidly when treatment is stopped. The risk of myopathy with fibrate treatment is increased in patients with renal impairment, and the risk of myopathy with statin treatment increases with co-administration of drugs that inhibit statin metabolism or transport. Other adverse effects are related to specific drugs, for example, clofibrate is associated with an increased risk of gallstones. Studies in elderly patients have not shown an increased risk of adverse effects with lipid-lowering drugs compared with younger patients, but in clinical practice there may be some increased risk, particularly with regards to drug interactions. Therefore, lipid-lowering drugs should be administered with extra caution to elderly patients.

Do statins cause cancer? A meta-analysis of large randomized clinical trials

PURPOSE

Although the short-term safety and tolerability of statins has been well established, their potential carcinogenicity in the long term is still debated. The goal of this study was to determine whether long-term treatment with statins is associated with an increased risk of fatal and nonfatal cancers.

METHODS

We searched the Medline database between January 1966 and December 1999 for randomized, controlled trials of human subjects in which monotherapy with a statin was compared with placebo. No language restrictions were applied. Only trials with a minimum treatment duration of 4 years and a minimum of 1,000 subjects were included. Studies that did not provide information on fatal or nonfatal cancers were excluded. Data on fatal and nonfatal cancers and all-cause mortality were extracted by a single nonblinded reviewer. Overall crude estimates of risk difference were computed by summing the numerators and denominators of trial-specific risk estimates.

RESULTS

Five trials met the inclusion criteria. The estimated differences in absolute risk between treatment and placebo were as follows (negative risks indicate that treatment was safer than placebo): all nonfatal cancers, 0.0% (95% confidence interval [CI]: -0.8% to 0.8%); all fatal cancers, -0.1% (95% CI: -0.7% to 0.4%); all fatal and nonfatal cancers combined, -0.1% (95% CI: -1.0% to 0.7%); and all-cause mortality, -1.5% (95% CI: 2.8% to 0.2%).

CONCLUSION

This study demonstrates no association between statin use over a 5-year period and the risk of fatal and nonfatal cancers. This conclusion is limited by the relatively short follow-up of the studies analyzed. Similar analyses of data from studies with longer follow-up periods would be valuable.

Apolipoprotein E genotype, serum lipids, and colorectal adenomas in Japanese men

We examined the relation of serum lipids and apolipoprotein E genotype to colorectal adenomas among 205 cases and 220 controls with normal colonoscopy in Japanese men. With adjustment for body mass index, cigarette smoking, alcohol use, and other covaiates, odds ratios of proximal and distal adenomas associated with the presence of an allele varepsilon4 were 0.59 (95% confidence interval 0.23-1.45) and 0.99 (0.50-1.98), respectively. While serum total and LDL cholesterol were unrelated to both proximal and distal adenomas, serum triglycerides were positively related to distal adenomas. The findings suggest that altered lipid metabolism may be differentially associated with tumorigenesis in the proximal and distal colorectum.

Association of low plasma cholesterol with mortality for cancer at various sites in men: 17-y follow-up of the prospective Basel study

BACKGROUND

Low serum cholesterol has been associated with an increased risk of cancer mortality in various studies, which has led to uncertainty regarding the benefit of lower blood cholesterol.

OBJECTIVE

The aim of our study was to evaluate the association between low blood cholesterol (<5.16 mmol/L) and cancer at sites that have rarely been evaluated. We placed special emphasis on the potential confounding effect of antioxidant vitamins.

DESIGN

Plasma concentrations of cholesterol and antioxidant vitamins were measured in 1971-1973 in 2974 men working in Basel, Switzerland. In 1990, the vital status of all participants was assessed.

RESULTS

Two hundred ninety of the participants had died from cancer, 87 from lung, 30 from prostate, 28 from stomach, and 22 from colon cancer. Group means for plasma cholesterol concentrations did not differ significantly between survivors and those who died from cancer at any of the studied sites. With plasma cholesterol, vitamins C and E, retinol, carotene, smoking, and age accounted for in a Cox model, an increase in total cancer mortality in lung, prostate, and colon but not in stomach cancer mortality was observed in men >60 y of age with low plasma cholesterol. When data from the first 2 y of follow-up were excluded from the analysis, the relative risk estimates remained practically unchanged with regard to lung cancer but decreased for colon, prostate, and overall cancer.

CONCLUSIONS

Increased cancer mortality risks associated with low plasma cholesterol were not explained by the confounding effect of antioxidant vitamins, but were attributed in part to the effect of preexisting cancer.

Novel role of oxidants in the molecular mechanism of action of peroxisome proliferators

Peroxisome proliferators are nongenotoxic rodent carcinogens that act as tumor promoters by increasing cell proliferation; however, their precise mechanism of action is not well understood. Oxidative DNA damage caused by leakage of hydrogen peroxide (H2O2) from peroxisomes was hypothesized initially as the mechanism by which these compounds cause liver tumors. It seems unlikely that oxidants of peroxisomal origin explain the mechanism of action of peroxisome proliferators because treatment with these compounds in vivo does not lead to increased H2O2 production. On the other hand, Kupffer cell-derived oxidants, such as superoxide, may play a role in initiating tumor nerosis factor-alpha (TNF-alpha) production that leads to hepatocyte proliferation. Peroxisome proliferators have been shown to activate Kupffer cells both in vitro and in vivo, and the use of Kupffer cell inhibitors such as methyl palmitate and dietary glycine have demonstrated that Kupffer cells are responsible for hepatocyte proliferation by mechanisms involve TNF-alpha. Moreover, peroxisome proliferators activate the transcription factor NF-kappaB, one of the major regulators of TNF-alpha expression, in Kupffer cells. Importantly, activation of NF-kappaB by peroxisome proliferators was shown to be oxidant-dependent, leading to the hypothesis that oxidants of Kupffer cell origin are involved in the mechanism of action. Many of the effects of peroxisome proliferators, including peroxisome induction and hepatomegaly, involve the peroxisome proliferator-activated receptor-alpha (PPARalpha). Recently, it was shown that peroxisome proliferator-induced cell proliferation and tumors require the PPARalpha. However, PPARalpha is not involved in TNF-alpha production by Kupffer cells because it is not expressed in this cell type. How it is involved in liver tumor remains unclear and one possible explanation is that both Kupffer cell TNF-alpha and parenchymal cell PPARalpha are required. Collectively, recent data are consistent with the hypothesis that oxidants play a role in signaling hepatocellular proliferation due to peroxisome proliferators via activation of NF-kappaB and incrase in mitogenic cytokines such as TNF-alpha.

Cancer and low levels of plasma cholesterol: the relevance of cholesterol precursors and products to incidence of cancer

BACKGROUND

It is proposed that the mechanism responsible for the low plasma cholesterol-cancer associations reported in the literature is related to blood and tissue levels of cholesterol precursors and products, reflecting rates of cholesterol synthesis rather than the plasma cholesterol level itself.

METHODS

Data have been collated from Japanese and Northern European studies on plasma levels of a cholesterol precursor, lathosterol, and one product, cholestanol, each a marker of cholesterol metabolism. Situations in which the rate of cholesterol synthesis is altered have also been examined for their relationship to cancer incidence.

RESULTS

The data though minimal suggest that lathosterol and cholestanol may be higher in the blood of the Japanese compared with the Northern Europeans, despite lower plasma cholesterol levels in the Japanese. In accord with the hypothesis the Japanese have a low incidence of many cancers. Cholesterol synthesis is lowered when dietary cholesterol and fat intake are increased and incidence of cancer is increased in these states. Conversely cholesterol synthesis is raised in vegetarianism, the Mediterranean diet, pregnancy, and lactation, and incidence of some cancers is lowered.

CONCLUSIONS

At least some of the variation in cancer incidence with plasma cholesterol levels and also with dietary saturated fat and cholesterol, as well as vegetarianism, the Mediterranean diet, pregnancy, and lactation, can be accounted for by their effects on the rate of cholesterol synthesis and the level of cholesterol precursors and/or products so generated.

[Cholesterolemia and pathology: update]

The relationship between plasma cholesterol levels and disease is more complex than presumed. If the pathogenic role of hypercholesterolemia in ischemic myocardial disease is now undoubted, the deleterious effect of hypocholesterolemia induced by diet or drugs, mainly on cancer mortality and violent death, is still controversial. In a prognostic point of view it seems that the degree of diminution of cholesterol levels is more important than hypocholesterolemia itself. Several hypothesis were formulated; however, hypocholesterolemia does not appear as a causal factor. A prudent approach of hypercholesterolemia remains desirable but without special alterations of the recent recommendations concerning the management of hypercholesterolemia.

Association between serum total cholesterol and HIV infection in a high-risk cohort of young men

Low serum total cholesterol (TC) is associated with a variety of nonatherosclerotic diseases, but the association of TC with infectious disease has been little studied. In this study, we examined the relationship between serum TC and HIV infection in members of a large health maintenance organization in Northern California. The cohort consisted of 2446 unmarried young men 15 to 49 years of age at high risk of HIV infection, defined as self-reported history of sexually transmitted disease or liver disease. Baseline measurements were taken between 1979 and 1985, and subjects were passively followed for HIV infection until the end of 1993 (average length of follow-up, 7.7 years). From a multivariate-adjusted Cox regression, the rate ratio (RR) of HIV infection was 1.66 (95% CI = 1.07, 2.56) for men with serum TC levels <160 mg/dl compared with those with TC levels between 160 and 199 mg/dl. Similar excess risk of AIDS and AIDS-related death was observed. These findings suggest that low serum TC levels should be considered a marker of increased risk of HIV infection in men already at heightened risk of HIV infection.

The Münster Heart Study (PROCAM): total mortality in middle-aged men is increased at low total and LDL cholesterol concentrations in smokers but not in nonsmokers

BACKGROUND

Some large epidemiological studies have shown an increase in mortality at low levels of total and LDL cholesterol. It has been speculated that low cholesterol levels may play a causative role in this association. To investigate this question, we analyzed all deaths occurring among middle-aged men in the Münster Heart Study (PROCAM), one of the largest prospective epidemiological studies of coronary heart disease risk markers in Europe.

METHODS AND RESULTS

In the Münster Heart Study, 10,856 men aged 36 to 65 years at study entry (46.8+/-7.3 years [mean+/-SD]) were followed for 4 to 14 years (7.1+/-2.4 years). During this period, 313 deaths occurred--46 from myocardial infarction, 48 from suspected or definite sudden cardiac death, 14 from cerebrovascular disease, and 10 from other diseases of the circulatory system. There were 121 deaths from cancer and 33 deaths from violent causes (injuries in 16, suicide in 14, and homicide in 3 cases). Death in 29 cases occurred from other causes and was unexplained in 12 cases. Total cholesterol, LDL cholesterol, and the LDL/HDL ratio showed a J-shaped relationship with total mortality. At high total and LDL cholesterol concentrations, increased mortality was due to increased coronary deaths. At low total and LDL cholesterol concentrations, increased mortality was seen in smokers only and was explained by an increase in smoking-related cancer deaths.

CONCLUSIONS

The increase in mortality at low levels of total and LDL cholesterol among middle-aged men in the Münster Heart Study is explained by an increase in smoking-related cancer deaths among smokers.

Prognostic value of cholesterol in women of different ages

We assessed the short-term and long-term prognostic relationship between cholesterol and mortality in women of different ages with the aid of statistical graphics. Our population-based cohort study involved 2873 women in the Framingham Heart Study, with a median follow-up of 31 years. The primary outcome was all-cause mortality. Secondary outcome measures were coronary heart disease, noncoronary heart disease, and stroke mortality. We found that significant age interactions were present in the relationships between total cholesterol and mortality from all causes, coronary heart disease (CHD), stroke, and non-CHD causes. For women ages < or = 55, cholesterol is related positively to both short-term (p > 0.05) and long-term (p = 0.05) all-cause mortality. For women ages 56-70, there are significant U-shaped relationships between cholesterol and both short-term and long-term all-cause mortality (p < 0.01). Lowest short-term and long-term mortality rates for women in this age group are at cholesterol values between 240 and 280 mg/dl. For women ages > 70, cholesterol < 240 mg/dl is associated with increased short-term mortality (p < 0.01), and no significant long-term association was detected. These cholesterol/mortality relationships and age interactions can be explained by patterns of association between mortality and both high- and low-density lipoprotein cholesterol among women in the different age groups. These results do not support the hypothesis that cholesterol < 200 mg/dl leads to decreased mortality in women > 55 years old.

Blood lipid and lipoprotein levels and the risk of cancer of the colon and rectum. A prospective study of 62,173 Norwegian men and women

BACKGROUND

Concern has been raised that a low total serum cholesterol level, although beneficial for cardiovascular diseases, may increase the risk of cancer. This prospective cohort study analyses the hypotheses that a low total serum cholesterol level or its subfractions (serum low-density-lipoprotein cholesterol, high-density-lipoprotein cholesterol, and triglycerides) increase the risk of cancer of the colon and rectum.

METHODS

Between 1977 and 1983, 62,173 men and women attended a health screening carried out by the Norwegian National Health Screening Service. The screening consisted of a questionnaire, anthropometric measurements, and samples of non-fasting blood drawn for analyses of serum total cholesterol, low-density-lipoprotein cholesterol, high-density-lipoprotein cholesterol, and triglycerides.

RESULTS

During the 7- to 13-year follow-up, 186 patients were found to have colon cancer and 106 rectal cancer by linkage to the Norwegian Cancer Registry. Among men there were no associations between blood lipid and lipoprotein levels and risk of cancer of the proximal colon, distal colon, or the rectum. Among women there was a formal statistically significant inverse relationship between level of total cholesterol and low-density-lipoprotein cholesterol and risk of distal colon cancer, and a positive trend between total cholesterol level and rectal cancer.

CONCLUSIONS

The statistically significant results among women were interpreted as incidental, and we conclude that blood lipid and lipoprotein levels were not associated with the risk of colon or rectum cancer in men or women in this cohort.

Low serum cholesterol and mortality. Which is the cause and which is the effect?

BACKGROUND

Many studies have reported an association between a low or lowered blood total cholesterol (TC) level and subsequent nonatherosclerotic disease incidence or death. The question of whether low TC is a true risk factor or alternatively a consequence of occult disease at the time of TC measurement remains unsettled. To shed new light onto this problem, we analyzed TC change over a 6- year period (from exam 1 in 1965 through 1968 to exam 3 in 1971 through 1974) in relation to subsequent 16-year mortality in a cohort of Japanese American men.

METHODS AND RESULTS

The study was based on 5941 men 45 to 68 years of age without prior history of coronary heart disease, stroke, cancer, or gastrointestinal-liver disease at exam 1 who also participated in exam 3 of the Honolulu Heart Program. The association of TC change with mortality end points was investigated with two different approaches (continuous and categorical TC change) with standard survival analysis techniques. Falling TC level was accompanied by a subsequent increased risk of death caused by some cancers (hemopoietic, esophageal, and prostate), noncardiovascular noncancer causes (particularly liver disease), and all causes. The risk-factor-adjusted rate of all-cause mortality was 30% higher (relative risk, 1.30; 95% CI, 1.06 to 1.59) among persons with a decline from middle (180 to 239 mg/dL) to low (< 180 mg/dL) TC than in persons remaining at a stable middle level. By contrast, there was no significant increase in all-cause mortality risk among cohort men with stable low TC levels. Nonillness mortality (deaths caused by trauma and suicide) was not related to either TC change or the average of TC levels in exams 1 and 3.

CONCLUSIONS

These results add strength to the reverse-causality proposition that catabolic diseases cause TC to decrease.

Gastric cancer among the Japanese in Hawaii

The incidence rate of gastric cancer among men of Japanese ancestry living in Hawaii is about one-third as high as that of their counterparts living in Japan. Because of this difference, a prospective study was conducted to identify factors related to the development of gastric cancer in Hawaii. Eight thousand and six (8,006) men born from 1900-1919 were examined from 1965 to 1968 and followed for over 25 years. During this time, 250 incident cases of gastric cancer were identified. The study has found the following: 1) prior infection with Helicobacter pylori bacteria increased the risk for stomach cancer; 2) cigarette smoking was positively associated with gastric cancer with age at which smoking started being an important risk factor; 3) after taking cigarette smoking into account, alcohol intake was not related to stomach cancer risk; 4) a low pepsinogen I level identified subjects at increased risk for the intestinal histologic type of gastric cancer; 5) a low serum ferritin level was a marker for increased risk of stomach cancer; 6) there was a weak indication that the intake of vegetables and fruits was inversely related to gastric cancer; 7) there was no association of stomach cancer with levels of serum cholesterol, serum uric acid, serum micronutrients (retinol, beta-carotene or alpha-tocopherol) or blood hematocrit; 8) there was also no association of gastric cancer with body mass index or physical activity.

Why sources of heterogeneity in meta-analysis should be investigated

Although meta-analysis is now well established as a method of reviewing evidence, an uncritical use of the technique can be very misleading. One common problem is the failure to investigate appropriately the sources of heterogeneity, in particular the clinical differences between the studies included. This paper distinguishes between the concepts of clinical and statistical heterogeneity and exemplifies the importance of investigating heterogeneity by using published meta-analyses of epidemiological studies of serum cholesterol concentration and clinical trials of its reduction. Although not without some dangers of speculative conclusions, prompted by overzealous inspection of the data to hand, a sensible investigation of sources of heterogeneity should increase both the scientific and the clinical relevance of the results of meta-analyses.

Assessing possible hazards of reducing serum cholesterol

OBJECTIVE

To assess whether low serum cholesterol concentration increases mortality from any cause.

DESIGN

Systematic review of published data on mortality from causes other than ischaemic heart disease derived from the 10 largest cohort studies, two international studies, and 28 randomised trials, supplemented by unpublished data on causes of death obtained when necessary.

MAIN OUTCOME MEASURES

Excess cause specific mortality associated with low or lowered serum cholesterol concentration.

RESULTS

The only cause of death attributable to low serum cholesterol concentration was haemorrhagic stroke. The excess risk was associated only with concentrations below about 5 mmol/l (relative risk 1.9, 95% confidence interval 1.4 to 2.5), affecting about 6% of people in Western populations. For noncirculatory causes of death there was a pronounced difference between cohort studies of employed men, likely to be healthy at recruitment, and cohort studies of subjects in community settings, necessarily including some with existing disease. The employed cohorts showed no excess mortality. The community cohorts showed associations between low cholesterol concentration and lung cancer, haemopoietic cancers, suicide, chronic bronchitis, and chronic liver and bowel disease; these were most satisfactorily explained by early disease or by factors that cause the disease lowering serum cholesterol concentration (depression causes suicide and lowers cholesterol concentration, for example). In the randomised trials nine deaths (from a total of 687 deaths not due to ischaemic heart disease in treated subjects) were attributed to known adverse effects of the specific treatments, but otherwise there was no evidence of an increased mortality from any cause arising from reduction in cholesterol concentration.

CONCLUSIONS

There is no evidence that low or reduced serum cholesterol concentration increases mortality from any cause other than haemorrhagic stroke. This risk affects only those people with a very low concentration and even in these will be outweighed by the benefits from the low risk of ischaemic heart disease.

Low cholesterol and risk of non-coronary mortality

Based on a systematic review of over 20 cohort studies, a clear association exists, for both men and women, between particularly low cholesterol levels and the rate of non-coronary mortality. The excess in women appears mainly confined to non-cancer causes, particularly respiratory and digestive diseases, while there is also an excess of deaths from cancer seen in men with low cholesterol levels. Higher mortality rates from trauma, haemorrhagic stroke and cirrhosis have also been observed. Much of this association is known to be as a consequence of the disease with a fall in cholesterol levels seen after developing a variety of inflammatory diseases. However, the excess risk of non-coronary heart disease deaths is still apparent by excluding deaths within five years suggesting that effect-cause is not the only explanation. Confounding still remains the most likely explanation for the association with an underlying chronic disease or risk factor causing both the low cholesterol and the fatal event. However, there is still the possibility that some of the increased risk is due to the low cholesterol. This makes it important that appropriately controlled trials of both drug and dietary interventions demonstrate net clinical benefit among those with low levels of coronary risk before cholesterol-lowering strategies are adopted more widely in these groups.

The relations between cervical cancer and serological markers of nutritional status

We evaluated whether differences in serological nutrient indicators between cases and controls were likely to be due to different usual levels for cases or to altered metabolism due to disease. Blood samples obtained as part of a case-control study of invasive cervical cancer conducted in Latin America were evaluated for case-control differences and for trends with stage of disease. Serum alpha- and beta-carotene, cryptoxanthin, and alpha- and gamma-tocopherol showed no trend with extent of disease, although Stage IV cases had lower alpha- and beta-carotene values than did other cases. A slight trend of decreasing values with stage was observed for serum retinol, lycopene, and lutein. For cholesterol and triglyceride concentrations, an inverse trend was observed with stage of disease, which suggested a clinical effect of the disease on blood lipids. Adjustment for smoking, alcohol intake, or oral contraceptive use did not alter observed relations, nor was there evidence that the altered blood nutrient levels differed by histological type. These data suggest that serum values for some carotenoids from Stage I, II, and III cervical cancer are suitable for etiological studies, but spurious results may be obtained if late-stage cases are included. Evidence of trends with severity of disease for cholesterol and triglycerides, and possibly for retinol, lycopene, and lutein, suggest that special attention be given to disease effects of these nutrients in studies of cervical cancer.

Serum lipids and left-sided adenomas of the large bowel: an extended study of self-defense officials in Japan

In the on-going study of men retiring from the Self-Defense Forces in Japan, we previously reported that serum total cholesterol was not related to colorectal adenomas but that men with low levels of serum high-density lipoprotein (HDL) cholesterol had an elevated adenoma risk. We examined whether the previous observation was reproducible in a different set of data accrued subsequently in the study. Serum total cholesterol, HDL-cholesterol, low-density lipoprotein (LDL) cholesterol, and triglycerides were compared between 138 cases of colorectal adenomas at the depth of 60 cm or less from the anus and 909 controls with normal sigmoidoscopy in the period from October 1988 to December 1990. There was virtually no relation between adenoma risk and any of the serum lipids studied with or without adjustment for smoking, alcohol use, and body mass index. In the analysis combining the earlier and present data, however, men with large adenomas (> or = 10 mm, n = 25) tended to have lower levels of total cholesterol and LDL-cholesterol compared with controls (n = 1,612); adjusted mean differences were -0.21 mmol/l (P = 0.24) and -0.26 mmol/l (P = 0.13), respectively. These findings are inconclusive, but hypocholesterolemia may be associated with the growth of colorectal adenoma.

Report of the Conference on Low Blood Cholesterol: Mortality Associations

BACKGROUND

A National Heart, Lung, and Blood Institute (NHLBI) Conference was held October 9-10, 1990, to review and discuss existing data on U-shaped relations found between mortality rates and blood total cholesterol levels (TC) in some but not other studies. Presentations were given from 19 cohort studies from the United States, Europe, Israel, and Japan. A representative of each study presented its findings and also submitted tables of proportional hazards regression coefficients for entry TC levels in regard to death, and these were incorporated into a formal statistical overview adjusted for age, diastolic blood pressure, cigarette smoking, body mass index, and alcohol intake, as available.

METHODS AND RESULTS

The U-shape for total mortality in men and the flat relation in women resulted largely from a positive relation of TC with coronary heart disease death and an inverse relation with deaths caused by some cancers (e.g., lung but not colon), respiratory disease, digestive disease, trauma, and residual deaths. Risk for combined noncardiovascular, noncancer causes of death decreased steadily across the range of TC. The conference considered possible explanations for the statistical associations found between low TC levels or active TC lowering and certain causes of death. One is that TC is lowered by some disease conditions themselves, such as wasting in chronic pulmonary disease or reduced production and secretion of cholesterol-bearing lipoproteins with liver disease. In this sort of situation, the TC:mortality association found in observational studies may be due to preexisting disease. This was addressed by excluding early deaths from the analysis, which did not change the results. The conference considered as well the biological function of cholesterol, which, if seriously deranged, might hypothetically cause a wide variety of diseases and dysfunction. The conference also considered the biological functions that might provide plausible mechanisms for the associations found.

CONCLUSIONS

Definitive interpretation of the associations observed was not possible, although most participants considered it likely that many of the statistical associations of low or lowered TC level are explainable by confounding in one form or another. The conference focused on the apparent existence and nature of these associations and on the need to understand their source rather than on any pertinence of the findings for public health policy. Further research is recommended to explain the observed associations of low TC levels (and TC lowering) with certain noncardiovascular diseases. This includes studies of the time course of TC change in disease, the relation of TC to morbidity, further studies of possible epidemiological confounding, monitoring of population trends in TC and mortality, further studies of the relations in women, auditing of noncardiovascular events in trials, studies of cell membrane, genetic and molecular links to cholesterol metabolism, TC level and disease, studies of disease manifestations in specific lipid disorders, and further study of the proposed causal mechanisms linking low TC and hemorrhagic stroke.

High protein, saturated fat and cholesterol diet, and low levels of serum lipids in colorectal cancer

In a case-control study probing the role of diet on the occurrence of colorectal cancer and undertaken in Athens, Greece, sera were collected from 100 cases and 100 controls, and serum total cholesterol, high-density-lipoprotein (HDL) cholesterol and triglycerides were determined. The biochemical results were analyzed in conjunction with nutrient intakes and a dietary score that summarizes in a linear way the dietary contrast between high-risk (high protein, saturated fat and dietary cholesterol; low vegetable) and low-risk (low protein, saturated fat and cholesterol; high vegetable) patterns. Cases with colorectal cancer had significantly (p less than 0.001) and substantially lower values of serum total cholesterol and particularly HDL cholesterol, but these associations did not reflect dietary practices, since protein intake and, to a lesser (and nonsignificant) extent, saturated fat and dietary cholesterol intake were higher among cases than among controls. In absolute terms, the dietary effect (as summarized in the linear dietary score) is more evident among persons with low serum total cholesterol and HDL cholesterol than among those with high levels of these serum lipids. These results indicate that a diet beneficial with respect to the risk of coronary heart disease is also likely to reduce the risk of colorectal cancer, even though low levels of serum total cholesterol and particularly HDL cholesterol represent important independent correlates of clinically overt colorectal cancer.

Can meta-analyses be trusted?

The enthusiasm for meta-analyses (or overviews) expressed by their proponents is not always shared by the broader medical community. To encourage constructive debate, we adopt a critical perspective on the conduct and interpretation of meta-analysis. We focus particularly on some of the statistical issues, especially heterogeneity between studies, and also on the extrapolation of meta-analysis findings to clinical practice. We conclude that meta-analysis is not an exact statistical science that provides definitive simple answers to complex clinical problems. It is more appropriately viewed as a valuable objective descriptive technique, which often furnishes clear qualitative conclusions about broad treatment policies, but whose quantitative results have to be interpreted cautiously.