Causal mediation analysis in nested case-control studies using conditional logistic regression.

The paper proposes an approach to causal mediation analysis in nested case-control study designs, often incorporated with countermatching schemes using conditional likelihood, and we compare the method's performance to that of mediation analysis using the Cox model for the full cohort with a continuous or dichotomous mediator. Simulation studies are conducted to assess our proposed method and investigate the efficiency relative to the cohort. We illustrate the method using actual data from two studies of potential mediation of radiation risk conducted within the Adult Health Study cohort of atomic-bomb survivors. The performance becomes comparable to that based on the full cohort, illustrating the potential for valid mediation analysis based on the reduced data obtained through the nested case-control design.

A Bayesian Semiparametric Model for Radiation Dose-Response Estimation.

In evaluating the risk of exposure to health hazards, characterizing the dose-response relationship and estimating acceptable exposure levels are the primary goals. In analyses of health risks associated with exposure to ionizing radiation, while there is a clear agreement that moderate to high radiation doses cause harmful effects in humans, little has been known about the possible biological effects at low doses, for example, below 0.1 Gy, which is the dose range relevant to most radiation exposures of concern today. A conventional approach to radiation dose-response estimation based on simple parametric forms, such as the linear nonthreshold model, can be misleading in evaluating the risk and, in particular, its uncertainty at low doses. As an alternative approach, we consider a Bayesian semiparametric model that has a connected piece-wise-linear dose-response function with prior distributions having an autoregressive structure among the random slope coefficients defined over closely spaced dose categories. With a simulation study and application to analysis of cancer incidence data among Japanese atomic bomb survivors, we show that this approach can produce smooth and flexible dose-response estimation while reasonably handling the risk uncertainty at low doses and elsewhere. With relatively few assumptions and modeling options to be made by the analyst, the method can be particularly useful in assessing risks associated with low-dose radiation exposures.

Serum interleukin-6 associated with hepatocellular carcinoma risk: a nested case-control study.

Inflammatory markers have been associated with increased risk of several cancers, including colon, lung, breast and liver, but the evidence is inconsistent. We conducted a nested case-control study in the longitudinal cohort of atomic-bomb survivors. The study included 224 hepatocellular carcinoma (HCC) cases and 644 controls individually matched to cases on gender, age, city and time and method of serum storage, and countermatched on radiation dose. We measured C-reactive protein (CRP) and interleukin (IL)-6 using stored sera obtained within 6 years before HCC diagnosis from 188 HCC cases and 605 controls with adequate volumes of donated blood. Analyses with adjustment for hepatitis virus infection, alcohol consumption, smoking habit, body mass index (BMI) and radiation dose showed that relative risk (RR) of HCC [95% confidence interval (CI)] in the highest tertile of CRP levels was 1.94 (0.72-5.51) compared to the lowest tertile (p = 0.20). RR of HCC (95% CI) in the highest tertile of IL-6 levels was 5.12 (1.54-20.1) compared to the lowest tertile (p = 0.007). Among subjects with BMI > 25.0 kg/m(2) , a stronger association was found between a 1-standard deviation (SD) increase in log IL-6 and HCC risk compared to subjects in the middle quintile of BMI (21.3-22.9 kg/m(2) ), resulting in adjusted RR (95% CI) of 3.09 (1.78-5.81; p = 0.015). The results indicate that higher serum levels of IL-6 are associated with increased HCC risk, independently of hepatitis virus infection, lifestyle-related factors and radiation exposure. The association is especially pronounced among subjects with obesity.

Radiation-dose response of glycophorin A somatic mutation in erythrocytes associated with gene polymorphisms of p53 binding protein 1.

Information on individual variations in response to ionizing radiation is still quite limited. Previous studies of atomic-bomb survivors revealed that somatic mutations at the glycophorin A (GPA) gene locus in erythrocytes were significantly elevated with radiation exposure dose, and that the dose response was significantly higher in survivors with subsequent cancer development compared to those without cancer development. Noteworthy in these studies were great inter-individual differences in GPA mutant fraction even in persons with similar radiation doses. It is hypothesized that persistent GPA mutations in erythrocytes of atomic-bomb survivors are derived from those in long-lived hematopoietic stem cell (HSC) populations, and that individual genetic backgrounds, specifically related to DNA double-strand break repair, contribute to individual differences in HSC mutability following radiation exposure. Thus, we examined the relationship between radiation exposure, GPA mutant fraction in erythrocytes, and single nucleotide polymorphisms (SNPs) of the key gene involved in DNA double-strand break repair, p53 binding protein 1 (53BP1). 53BP1 SNPs and inferred haplotypes demonstrated a significant interaction with radiation dose, suggesting that radiation-dose response of GPA somatic mutation is partly dependent on 53BP1 genotype. It is also possible that 53BP1 plays a significant role in DNA double-strand break repair in HSCs following radiation exposure.

Radiation risk of individual multifactorial diseases in offspring of the atomic-bomb survivors: a clinical health study.

There is no convincing evidence regarding radiation-induced heritable risks of adult-onset multifactorial diseases in humans, although it is important from the standpoint of protection and management of populations exposed to radiation. The objective of the present study was to examine whether parental exposure to atomic-bomb (A-bomb) radiation led to an increased risk of common polygenic, multifactorial diseases-hypertension, hypercholesterolaemia, diabetes mellitus, angina pectoris, myocardial infarction or stroke-in the first-generation (F1) offspring of A-bomb survivors. A total of 11,951 F1 offspring of survivors in Hiroshima or Nagasaki, conceived after the bombing, underwent health examinations to assess disease prevalence. We found no evidence that paternal or maternal A-bomb radiation dose, or the sum of their doses, was associated with an increased risk of any multifactorial diseases in either male or female offspring. None of the 18 radiation dose-response slopes, adjusted for other risk factors for the diseases, was statistically significantly elevated. However, the study population is still in mid-life (mean age 48.6 years), and will express much of its multifactorial disease incidence in the future, so ongoing longitudinal follow-up will provide increasingly informative risk estimates regarding hereditary genetic effects for incidence of adult-onset multifactorial disease.

Impact of radiation and hepatitis virus infection on risk of hepatocellular carcinoma.

UNLABELLED: In cohort studies of atomic bomb survivors and Mayak nuclear facility workers, radiation-associated increases in liver cancer risk were observed, but hepatitis B virus (HBV) and hepatitis C virus (HCV) infections were not taken strictly into account. We identified 359 hepatocellular carcinoma (HCC) cases between 1970 and 2002 in the cohort of atomic bomb survivors and estimated cumulative incidence of HCC by radiation dose. To investigate contributions of radiation exposure and hepatitis virus infection to HCC risk, we conducted a nested case-control study using sera stored before HCC diagnosis in the longitudinal cohort of atomic bomb survivors. The study included 224 HCC cases and 644 controls that were matched to the cases on gender, age, city, and time and method of serum storage, and countermatched on radiation dose. The cumulative incidence of HCC by follow-up time and age increased significantly with radiation dose. The relative risk (RR) of HCC for radiation at 1 Gy was 1.67 (95% confidence interval: 1.22-2.35) with adjustment for alcohol consumption, body mass index (BMI), and smoking habit, whereas the RRs for HBV or HCV infection alone were 63 (20-241) and 83 (36-231) with such adjustment, respectively. Those estimates changed little when radiation and hepatitis virus infection were fit simultaneously. The RR of non-B, non-C HCC at 1 Gy was 1.90 (1.02-3.92) without adjustment for alcohol consumption, BMI, or smoking habit and 2.74 (1.26-7.04) with such adjustment. CONCLUSION: These results indicate that radiation exposure and HBV and HCV infection are associated independently with increased HCC risk. In particular, radiation exposure was a significant risk factor for non-B, non-C HCC with no apparent confounding by alcohol consumption, BMI, or smoking habit.

Body iron stores and breast cancer risk in female atomic bomb survivors.

Iron can be a potent pro-oxidant and, on this basis, elevated body iron may increase the risk of cancer. Although epidemiological evidence is mixed, there is overall support for this possibility. In addition, because of this same oxidative capacity, body iron levels may alter radiation sensitivity. In the present study, a nested case-control study of breast cancer was conducted in Japanese atomic bomb survivors. Stored serum samples from the Adult Health Study cohort were assayed for ferritin levels and joint statistical analyses were conducted of ferritin and radiation dose on the risk of breast cancer. Serum ferritin is the best feasible indicator of body iron levels in otherwise healthy people. A total of 107 cases and 212 controls were available for analysis. The relative risk (RR) of breast cancer for a 1 log unit increase in ferritin was 1.4 (95% confidence interval 1.1-1.8). This translates to an RR of 1.64 comparing high and low values of the interquartile range among controls (58 and 13.2 ng/mL, respectively). The results support the hypothesis that elevated body iron stores increase the risk of breast cancer. However, the study was inconclusive regarding the question of whether body iron alters radiation-induced breast cancer risk.

Radiation Risk of Ovarian Cancer in Atomic Bomb Survivors: 1958-2009.

There is limited evidence concerning the association between radiation exposure and ovarian cancer. We evaluated radiation risk of ovarian cancer between 1958 and 2009 among 62,534 female atomic bomb survivors in the Life Span Study cohort, adding 11 years of follow-up from the previously reported study. Poisson regression methods were used to estimate excess relative risk per Gy (ERR/Gy) for total ovarian cancer and according to tumor type. We assessed the modifying effect of follow-up period and other factors on the radiation risk. We ascertained 288 first primary ovarian cancers including 77 type 1 epithelial cancers, 75 type 2 epithelial cancers, 66 epithelial cancers of undetermined type and 70 other cancers. Radiation dose was positively, although not significantly, associated with risk of total ovarian cancer [ERR/Gy = 0.30, 95% confidence interval (CI): -0.22 to 1.11]. There was a suggestion of heterogeneity in radiation effects (P = 0.08) for type 1 (ERR/Gy = -0.32, 95% CI: <-0.32 to 0.88) and type 2 cancers (ERR/Gy = 1.24, 95% CI: -0.08 to 4.16). There were no significant trends in the ERR with time since exposure or age at exposure. Further follow-up will help characterize more accurately the patterns of radiation risk for total ovarian cancer and its types.

Lung cancer susceptibility among atomic bomb survivors in relation to CA repeat number polymorphism of epidermal growth factor receptor gene and radiation dose.

Lung cancer is a leading cause of cancer death worldwide. Prevention could be improved by identifying susceptible individuals as well as improving understanding of interactions between genes and etiological environmental agents, including radiation exposure. The epidermal growth factor receptor (EGFR)-signaling pathway, regulating cellular radiation sensitivity, is an oncogenic cascade involved in lung cancer, especially adenocarcinoma. The cytosine adenine (CA) repeat number polymorphism in the first intron of EGFR has been shown to be inversely correlated with EGFR production. It is hypothesized that CA repeat number may modulate individual susceptibility to lung cancer. Thus, we carried out a case-cohort study within the Japanese atomic bomb (A-bomb) survivor cohort to evaluate a possible association of CA repeat polymorphism with lung cancer risk in radiation-exposed or negligibly exposed (<5 mGy) A-bomb survivors. First, by dividing study subjects into Short and Long genotypes, defined as the summed CA repeat number of two alleles < or = 37 and > or = 38, respectively, we found that the Short genotype was significantly associated with an increased risk of lung cancer, specifically adenocarcinoma, among negligibly exposed subjects. Next, we found that prior radiation exposure significantly enhanced lung cancer risk of survivors with the Long genotype, whereas the risk for the Short genotype did not show any significant increase with radiation dose, resulting in indistinguishable risks between these genotypes at a high radiation dose. Our findings imply that the EGFR pathway plays a crucial role in assessing individual susceptibility to lung adenocarcinoma in relation to radiation exposure.

Predicting future excess events in risk assessment.

Risk characterization in a study population relies on cases of disease or death that are causally related to the exposure under study. The number of such cases, so-called "excess" cases, is not just an indicator of the impact of the risk factor in the study population, but also an important determinant of statistical power for assessing aspects of risk such as age-time trends and susceptible subgroups. In determining how large a population to study and/or how long to follow a study population to accumulate sufficient excess cases, it is necessary to predict future risk. In this study, focusing on models involving excess risk with possible effect modification, we describe a method for predicting the expected magnitude of numbers of excess cases and assess the uncertainty in those predictions. We do this by extending Bayesian APC models for rate projection to include exposure-related excess risk with possible effect modification by, e.g., age at exposure and attained age. The method is illustrated using the follow-up study of Japanese Atomic-Bomb Survivors, one of the primary bases for determining long-term health effects of radiation exposure and assessment of risk for radiation protection purposes. Using models selected by a predictive-performance measure obtained on test data reserved for cross-validation, we project excess counts due to radiation exposure and lifetime risk measures (risk of exposure-induced deaths (REID) and loss of life expectancy (LLE)) associated with cancer and noncancer disease deaths in the A-Bomb survivor cohort.

Risk factors for hepatocellular carcinoma in a Japanese population: a nested case-control study.

BACKGROUND: Epidemiologic studies have shown effects of lifestyle-related factors on risk for hepatocellular carcinoma. However, few cohort studies have incorporated, in a strict and in-depth manner, hepatitis B virus (HBV) and hepatitis C virus (HCV) infections or investigated synergism between such factors. METHODS: We conducted a nested case-control study using sera stored before hepatocellular carcinoma diagnosis in the longitudinal cohort of atomic bomb survivors. The study included 224 hepatocellular carcinoma cases and 644 controls that were matched to the cases on gender, age, city, time of serum storage, and method of serum storage, and countermatched on radiation dose. RESULTS: Univariate analysis showed that HBV and HCV infections, alcohol consumption, smoking habit, body mass index (BMI), and diabetes mellitus were associated with increased hepatocellular carcinoma risk, whereas coffee drinking was associated with decreased hepatocellular carcinoma risk. Multivariate relative risks of hepatocellular carcinoma (95% confidence interval) were 45.8 (15.2-138), 101 (38.7-263), 70.7 (8.3-601), 4.36 (1.48-13.0), and 4.57 (1.85-11.3), for HBV infection alone, HCV infection alone, both HBV and HCV infections, alcohol consumption of > or =40 g of ethanol per day, and BMI of >25.0 kg/m(2) 10 years before diagnosis, respectively. HBV and HCV infection and BMI of >25.0 kg/m(2) remained independent risk factors even after adjusting for severity of liver fibrosis. Among HCV-infected individuals, the relative risk of hepatocellular carcinoma for a 1 kg/m(2) increase in BMI was 1.39 (P = 0.003). CONCLUSIONS: To limit the risk for hepatocellular carcinoma, control of excess weight may be crucial for individuals with chronic liver disease, especially those with chronic hepatitis C.

Allowance for random dose estimation errors in atomic bomb survivor studies: a revision.

Allowing for imprecision of radiation dose estimates for A-bomb survivors followed up by the Radiation Effects Research Foundation can be improved through recent statistical methodology. Since the entire RERF dosimetry system has recently been revised, it is timely to reconsider this. We have found that the dosimetry revision itself does not warrant changes in these methods but that the new methodology does. In addition to assumptions regarding the form and magnitude of dose estimation errors, previous and current methods involve the apparent distribution of true doses in the cohort. New formulas give results conveniently and explicitly in terms of these inputs. Further, it is now possible to use assumptions about two components of the dose errors, referred to in the statistical literature as "classical" and "Berkson-type". There are indirect statistical indications, involving non-cancer biological effects, that errors may be somewhat larger than assumed before, in line with recommendations made here. Inevitably, methods must rely on uncertain assumptions about the magnitude of dose errors, and it is comforting to find that, within the range of plausibility, eventual cancer risk estimates are not very sensitive to these.

White blood cell count, especially neutrophil count, as a predictor of hypertension in a Japanese population.

Although several studies have shown that high WBC count is a risk factor for hypertension, the relationship between WBC count and the incidence of hypertension in Japanese is poorly understood, as are the effects of WBC components on that relationship. Our objective was to verify in a Japanese population whether WBC or differential WBC count predicts hypertension incidence. A total of 9,383 initially hypertension-free subjects (3,356 men and 6,027 women), whose WBC counts were within the normal range (3,000 to < 10,000 cells/mm3), were followed from 1965 to 2004. During this 40-year follow-up, 4,606 subjects developed hypertension. After adjusting for conventional risk factors, including smoking status, we found that elevated WBC count was associated with hypertension incidence in a Cox regression model with both fixed and time-varying covariates for women. For men, elevated WBC count was a significant risk factor for hypertension only in the time-varying Cox-regression covariate. We also observed a significant association between increased neutrophil count and hypertension incidence among women. In a fully adjusted model, the relative risks of hypertension incidence, from the lowest to the highest quartiles of neutrophil count, were 1.00, 1.18, 1.28, and 1.22 in women (p for trend < 0.001). In conclusion, elevated WBC count predicted an increased incidence of hypertension in Japanese, especially among females. Moreover, neutrophils were the major WBC component contributing to the increased risk.

Radiation Risks of Uterine Cancer in Atomic Bomb Survivors: 1958-2009.

BACKGROUND: Ionizing radiation is known to be capable of causing cancer of many organs, but its relationship with uterine cancer has not been well characterized. METHODS: We studied incidence of uterine cancer during 1958-2009 among 62 534 female atomic bomb survivors. Using Poisson regression analysis, we fitted excess relative risk (ERR) models to uterine cancer rates adjusted for several lifestyle and reproductive factors. Person-years at risk were also adjusted for the probability of prior hysterectomy, because it could affect the subsequent risk of uterine cancer. We assessed the modifying effect of age and other factors on the radiation risk. For analysis of the modifying effect of age at radiation exposure around menarche, we compared the radiation risk for several exposure-age categories as well as using parametric models. RESULTS: There were 224 uterine corpus cancers and 982 cervical cancers. We found a significant association between radiation dose and risk of corpus cancer (ERR per Gray [ERR/Gy] = 0.73, 95% confidence interval [CI] = 0.03 to 1.87) but not for cervical cancer (ERR/Gy = 0.00, 95% CI = -0.22 to 0.31). For corpus cancer, we found statistically significant heterogeneity in ERR/Gy by age (P heterogeneity = .001) with elevated risk for women exposed to radiation between ages 11 and 15 years (ERR/Gy = 4.10, 95% CI = 1.47 to 8.42) and no indication of a radiation effect for exposures before or after this exposure-age range. CONCLUSIONS: The current data suggest that uterine corpus is especially sensitive to the carcinogenic effect of radiation exposure occurring during the mid-pubertal period preceding menarche. There is little evidence for a radiation effect on cervical cancer risk.

Bioavailable serum estradiol may alter radiation risk of postmenopausal breast cancer: a nested case-control study.

PURPOSE: Ionizing radiation and high levels of circulating estradiol are known breast cancer carcinogens. We investigated the risk of first primary postmenopausal breast cancer in relation to the combined effects of whole-body ionizing radiation exposure and prediagnostic levels of postmenopausal sex hormones, particularly bioavailable estradiol (bE2). MATERIALS AND METHODS: A nested case-control study of 57 incident breast cancer cases matched with 110 controls among atomic bomb survivors. Joint effects of breast radiation dose and circulating levels of sex hormones were assessed using binary regression and path analysis. RESULTS AND CONCLUSION: Radiation exposure, higher levels of bE2, testosterone and progesterone, and established reproductive risk factors were positively associated with postmenopausal breast cancer risk. A test for mediation of the effect of radiation via bE2 level suggested a small (14%) but significant mediation (p = 0.004). The estimated interaction between radiation and bE2 was large but not significant (interaction = 3.86; p = 0.32). There is accumulating evidence that ionizing radiation not only damages DNA but also alters other organ systems. While caution is needed, some portion of the radiation risk of postmenopausal breast cancer appeared to be mediated through bE2 levels, which may be evidence for cancer risks due to both direct and indirect effects of radiation.

Individual variation of somatic gene mutability in relation to cancer susceptibility: prospective study on erythrocyte glycophorin a gene mutations of atomic bomb survivors.

It has previously been reported that hemizygous mutant fraction (Mf) at the glycophorin A (GPA) locus in erythrocytes increased with radiation dose in heterozygotes among Hiroshima and Nagasaki atomic bomb survivors. In the present study, we analyzed the relationship between GPA Mf and cancer risk using newly developed cancers among previously cancer-free subjects whose GPA Mf had been measured between 1988 and 1996. Among 1,723 survivors (1,117 in Hiroshima and 606 in Nagasaki), we identified 186 subjects who developed a first cancer by the end of 2000. We compared the radiation dose responses of GPA Mf between cancer and cancer-free groups using a linear-quadratic model fit by multiple regression analysis in combination with age, sex, and city. The slope of the GPA Mf dose-response curve was significantly higher in the cancer group than in the cancer-free group among Hiroshima subjects. Moreover, no significant difference of GPA Mf between cancer and cancer-free groups was found in unexposed controls in the two cities. The same conclusions were obtained using a linear dose-response model and by further analysis using Cox regression of cancer incidence. These findings suggest that there might be interindividual variation in mutability of somatic genes and that Hiroshima survivors who have higher mutability in response to radiation exposure would be expected to have a higher probability of suffering radiation-related cancer.

Aging-related changes in human T-cell repertoire over 20years delineated by deep sequencing of peripheral T-cell receptors.

Recent deep sequencing studies on T-cell receptor (TCR) repertoire have provided robust data to characterize diversity of T-cell immune responsiveness to a wide variety of peptide antigens, including viral and tumor antigens. The human TCR repertoire declines with age, but this decline has not been fully investigated longitudinally in individuals. Using a deep sequencing approach, we analyzed TCRß repertoires longitudinally over approximately 20years, with ages ranging from 23 to 50years at the start (23 to 65years overall), in peripheral-blood CD4 and CD8 T-cell populations that were collected and cryopreserved 3 times at intervals of approximately 10years from each of 6 healthy adults (3 men and 3 women). Sequence data at the hypervariable complementarity determining region 3 (CDR3) in the TCRB gene locus were evaluated by applying a random-coefficient statistical regression model. Two outcomes were analyzed: total number of distinct TCRB CDR3 sequences as a TCR diversity metric, and clonality of the T-cell populations. TCR repertoire diversity decreased (p<0.001) and frequencies of clonal populations increased (p=0.003) with age in CD8 T cells, whereas CD4 T cells retained fairly diverse TCR repertoires along with relatively low clonality. We also found that approximately 10-30% and 30-80% of read sequences in CD4 and CD8 T cells, respectively, overlapped at different ages within each individual, indicating long-term stable maintenance of T-cell clonal composition. Moreover, many of the most frequent TCRB CDR3 sequences (i.e., top T-cell clones) persisted over 20years, and some of them expanded and exerted a dominating influence on clonality of peripheral T-cell populations. It is thus possible that persistence or expansion of top T-cell clones is a driver of T-cell immunity aging, and therefore represents a potential interventional target.

CD14 and IL18 gene polymorphisms associated with colorectal cancer subsite risks among atomic bomb survivors.

Colorectal cancer (CRC) is a common malignancy worldwide, and chronic inflammation is a risk factor for CRC. In this study, we carried out a cohort study among the Japanese atomic bomb (A-bomb) survivor population to investigate any association between immune- and inflammation-related gene polymorphisms and CRC. We examined the effects of six single-nucleotide polymorphisms of CD14 and IL18 on relative risks (RRs) of CRC. Results showed that RRs of CRC, overall and by anatomic subsite, significantly increased with increasing radiation dose. The CD14-911A/A genotype showed statistically significant higher risks for all CRC and distal CRC compared with the other two genotypes. In addition, the IL18-137 G/G genotype showed statistically significant higher risks for proximal colon cancer compared with the other two genotypes. In phenotype-genotype analyses, the CD14-911A/A genotype presented significantly higher levels of membrane and soluble CD14 compared with the other two genotypes, and the IL18-137 G/G genotype tended to be lower levels of plasma interleukin (IL)-18 compared with the other two genotypes. These results suggest the potential involvement of a CD14-mediated inflammatory response in the development of distal CRC and an IL18-mediated inflammatory response in the development of proximal colon cancer among A-bomb survivors.

Improving the efficiency of nested case-control studies of interaction by selecting controls using counter matching on exposure.

BACKGROUND: Studies of the effect of exposure to a risk factor measured in an entire cohort may be augmented by nested case-control subsets to investigate confounding or effect modification by additional factors not practically assessed on all cohort members. We compared three control-selection strategies-matching on exposure, counter matching on exposure, and random sampling-to determine which was most efficient in a situation where exposure is a known, continuous variable and high doses are rare. METHODS: We estimated the power to detect interaction using four control-to-case ratios (1:1, 2:1, 4:1, and 8:1) in a planned case-control study of the joint effect of atomic bomb radiation exposure and serum oestradiol levels on breast cancer. Radiation dose is measured in the entire cohort, but because neither serum oestradiol level nor the true degree of interaction was known, we simulated values of oestradiol and hypothetical levels of oestradiol-radiation interaction. RESULTS: Compared with random sampling, power to detect interaction was similarly higher with either matching or counter matching with two or more controls. CONCLUSIONS: Because counter matching is generally at least as efficient as random sampling, whereas matching on exposure can result in loss of efficiency and precludes estimation of exposure risk, we recommend counter matching for selecting controls in nested case-control studies of the joint effects of multiple risk factors when one is previously measured in the full cohort.