Trial
ASPREE and ASPREE-XT: aspirin in older adults
Research Observatory
Evidence Update
Current evidence
ASPREE compared enteric-coated aspirin 100 mg daily with placebo in 19,114 independent older adults in Australia and the US, mostly aged 70 or older (specified US minority groups from age 65), without overt cardiovascular disease, dementia or disability limiting independence. Median trial follow-up was 4.7 years.The main combined outcome rate was 21.5 with aspirin and 21.2 with placebo per 1,000 person-years. Aspirin did not significantly improve this outcome during the trial. The selected, independent population and finite follow-up limit broader conclusions. Taking the assigned tablets declined over time. Major bleeding is an important separate result in the same trial, shown alongside this finding.Cardiovascular event rates were 10.7 with aspirin and 11.3 with placebo per 1,000 person-years, without a statistically clear reduction. Major bleeding rates were 8.6 and 6.2 per 1,000 person-years, respectively. Major bleeding was more frequent with aspirin. The reported rate difference is 2.4 additional major bleeds per 1,000 person-years. This rate difference is not the percentage of people who bleed over the whole trial.Death rates were 12.7 with aspirin and 11.1 with placebo per 1,000 person-years. Cancer-related deaths contributed to the imbalance. The mortality imbalance is important, but it was a secondary finding. Looking separately at causes of death after the fact and testing multiple outcomes limit broad conclusions about cancer.Extended follow-up did not demonstrate a statistically significant reduction in overall MACE by original aspirin assignment. The whole-period and post-trial-only estimates describe different periods. Whole-period major bleeding remained an important adverse finding. Selection into follow-up and changes in treatment exposure limit interpretation of the post-trial estimates.Original aspirin assignment did not show a significant reduction in cancer incidence. Cancer mortality was higher overall; the post-trial-only mortality estimate remained uncertain. Later treatment changed, and the post-trial participants were selected from cancer-free survivors. Those differences limit comparisons across the two time periods.ASPREE-XT is follow-up of this parent trial, not another randomized trial. Later medication exposure was observational. The original primary outcome, secondary outcomes, whole-period results and post-trial-only results must be read separately.
What changed
Study findings and context:In ASPREE older adults without overt cardiovascular disease, dementia or independence-limiting disability, assigned low-dose aspirin did not significantly improve the primary disability-free survival composite over the observed trial period.ASPREE reported no significant reduction in secondary cardiovascular events with aspirin, while major hemorrhage was more frequent; the absolute event-rate differences and relative hazard ratios describe different scales.ASPREE observed a higher secondary all-cause mortality rate in the aspirin group, with cancer-related deaths contributing to the imbalance; this unexpected observation should not be presented as established universal cancer causation.ASPREE extended follow-up did not demonstrate a statistically significant reduction in overall MACE by original aspirin assignment; post-trial estimates and whole-period hemorrhage must be kept separate from continuously randomized treatment effects.Over median 8.6-year ASPREE/ASPREE-XT follow-up, original low-dose aspirin assignment did not show a significant reduction in cancer incidence; higher overall cancer mortality and the post-trial-only uncertain mortality estimate refer to different periods and selected populations.
What did not change
The combined result does not prove zero effect on every individual part of the outcome. It does not answer whether someone already taking aspirin should stop.The findings do not supply an individual prediction of benefit or bleeding. They do not apply directly to people using aspirin after established cardiovascular disease.This unexpected observation does not establish universal cancer causation or show that stopping aspirin reverses a cancer risk.Later medication use was not continuously randomized. These results do not establish a causal effect of stopping aspirin or a rebound after stopping.These different periods and selected populations do not establish an effect of continuous long-term aspirin use or a causal effect of stopping it.
Why it matters
Makes the trial’s main outcome visible before secondary findings or claims about longer independent life. Shows the uncertain cardiovascular result together with the important bleeding harm. Keeps an important adverse finding visible without turning it into a general cancer-causation claim. Separates the original treatment assignment from what people took later and keeps analysis periods distinct. Helps readers distinguish new diagnoses from deaths and whole-period findings from selected post-trial results.
Main limitation
Healthy independent baseline and finite follow-up limit extrapolation to frailty, established CVD or lifelong use. Adherence declined; trial assignment is not continuous exposure. Composite null is not equivalence, component-specific null or a direct stopping-trial result. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.No individual benefit/bleeding prediction or personalized starting/stopping recommendation. Secondary prevention and other doses/populations are not tested here. Source-qualified rate differences do not supply cumulative risk, NNT/NNH, equal severity or long-term lifetime benefit. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.Unexpected signal, secondary/exploratory hierarchy and multiple outcomes limit a broad causal cancer headline. Same participants as C05/C06, not replication. No inference that aspirin causes cancer in every age group or that stopping reverses a cancer risk. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.Consent/survivor selection, in-trial-event exclusions and changing exposure. MACE differs from the original broader cardiovascular endpoint; do not pool HRs blindly. Abstract does not supply per-arm event counts or all absolute risks; none are invented. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.Post-trial treatment changes and cancer-free survivor/consent selection; time windows and denominators differ. Per-arm absolute cancer risks are not qualified from the abstract. PMCID is a locator; index release date is 2027, so current free full-text availability is not assumed. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.
Next proof
Use the linked original reports to check the findings. Full methods, tables and supplements were not reviewed in this evidence package. Broader clinical questions require separate evidence review.
Study findings and exact evidence
Aspirin 100 mg daily (ASPREE randomized assignment)
Aspirin did not significantly improve the main independence-related outcome
ACCEPT
What was tested? Enteric-coated aspirin 100 mg daily was compared with placebo. The main outcome combined death, dementia and persistent physical disability.
Who participated? ASPREE enrolled 19,114 independent older adults in Australia and the US, mostly aged 70 or older. Specified US minority groups could enter from age 65. Participants had no overt cardiovascular disease, dementia or disability limiting independence. Median trial follow-up was 4.7 years.
Findings The main combined outcome rate was 21.5 with aspirin and 21.2 with placebo per 1,000 person-years. Aspirin did not significantly improve this outcome during the trial.
What was not demonstrated? The combined result does not prove zero effect on every individual part of the outcome. It does not answer whether someone already taking aspirin should stop.
Harms and limitations The selected, independent population and finite follow-up limit broader conclusions. Taking the assigned tablets declined over time. Major bleeding is an important separate result in the same trial, shown alongside this finding.
In ASPREE older adults without overt cardiovascular disease, dementia or independence-limiting disability, assigned low-dose aspirin did not significantly improve the primary disability-free survival composite over the observed trial period.
Exact accepted support, hierarchy and limitations
exact accepted proposition
In ASPREE older adults without overt cardiovascular disease, dementia or independence-limiting disability, assigned low-dose aspirin did not significantly improve the primary disability-free survival composite over the observed trial period.
exact numeric support
[{"absolute_difference_method": "LN arithmetic: 21.5 minus 21.2; no absolute CI derived", "absolute_difference_unit": "events per 1000 person-years", "absolute_rate_difference": 0.3, "confidence_interval": [0.92, 1.11], "confidence_level": 0.95, "denominator_comparator": 9589, "denominator_intervention": 9525, "effect_scale": "hazard_ratio", "endpoint": "Death/dementia/persistent physical disability", "estimate": 1.01, "p_value": 0.79, "rate_comparator": 21.2, "rate_intervention": 21.5, "rate_unit": "per 1000 person-years"}]exact original limitations
Healthy independent baseline and finite follow-up limit extrapolation to frailty, established CVD or lifelong use. Adherence declined; trial assignment is not continuous exposure. Composite null is not equivalence, component-specific null or a direct stopping-trial result. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.
exposure and comparator
Enteric-coated aspirin 100 mg daily Comparator: Matching placebo; randomized double-blind trial
follow up
Median 4.7 years; primary-endpoint futility stop
interpretation
Hazard ratios compare event hazards during follow-up; they are not individual probabilities or cumulative risk differences. Confidence intervals express uncertainty. Person-years add participants’ follow-up time.
outcome hierarchy
Primary: death from any cause, dementia or persistent physical disability. Components and major hemorrhage are secondary.
population
19114 independent older adults in Australia/US; >=70, or >=65 for specified US minority populations; median age 74; 56.4% women. Species: Homo sapiens.
Cardiovascular benefit was uncertain; major bleeding increased
ACCEPT
What was tested? In the same aspirin-versus-placebo trial, researchers assessed cardiovascular events and major bleeding as secondary outcomes. The cardiovascular outcome combined coronary death, nonfatal heart attack, stroke and heart-failure hospital admission.
Who participated? These results come from the same 19,114 ASPREE participants, with median follow-up of 4.7 years. They are not another independent trial.
Findings Cardiovascular event rates were 10.7 with aspirin and 11.3 with placebo per 1,000 person-years, without a statistically clear reduction. Major bleeding rates were 8.6 and 6.2 per 1,000 person-years, respectively.
What was not demonstrated? The findings do not supply an individual prediction of benefit or bleeding. They do not apply directly to people using aspirin after established cardiovascular disease.
Harms and limitations Major bleeding was more frequent with aspirin. The reported rate difference is 2.4 additional major bleeds per 1,000 person-years. This rate difference is not the percentage of people who bleed over the whole trial.
ASPREE reported no significant reduction in secondary cardiovascular events with aspirin, while major hemorrhage was more frequent; the absolute event-rate differences and relative hazard ratios describe different scales.
Exact accepted support, hierarchy and limitations
exact accepted proposition
ASPREE reported no significant reduction in secondary cardiovascular events with aspirin, while major hemorrhage was more frequent; the absolute event-rate differences and relative hazard ratios describe different scales.
exact numeric support
[{"absolute_difference_method": "LN arithmetic: 10.7 minus 11.3; not a cumulative risk difference", "absolute_difference_unit": "events per 1000 person-years", "absolute_rate_difference": -0.6, "confidence_interval": [0.83, 1.08], "confidence_level": 0.95, "denominator_comparator": 9589, "denominator_intervention": 9525, "effect_scale": "hazard_ratio", "endpoint": "Fatal CHD, nonfatal MI, fatal/nonfatal stroke or HF hospitalization", "estimate": 0.95, "rate_comparator": 11.3, "rate_intervention": 10.7, "rate_unit": "per 1000 person-years"}, {"absolute_difference_method": "LN arithmetic: 8.6 minus 6.2; no NNH derived", "absolute_difference_unit": "events per 1000 person-years", "absolute_rate_difference": 2.4, "confidence_interval": [1.18, 1.62], "confidence_level": 0.95, "denominator_comparator": 9589, "denominator_intervention": 9525, "effect_scale": "hazard_ratio", "endpoint": "Major hemorrhage", "estimate": 1.38, "p_value": "<0.001", "rate_comparator": 6.2, "rate_intervention": 8.6, "rate_unit": "per 1000 person-years"}]exact original limitations
No individual benefit/bleeding prediction or personalized starting/stopping recommendation. Secondary prevention and other doses/populations are not tested here. Source-qualified rate differences do not supply cumulative risk, NNT/NNH, equal severity or long-term lifetime benefit. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.
exposure and comparator
Enteric-coated aspirin 100 mg daily Comparator: Placebo, same ASPREE randomization
follow up
Median 4.7 years
interpretation
Hazard ratios compare event hazards during follow-up; they are not individual probabilities or cumulative risk differences. Confidence intervals express uncertainty. Person-years add participants’ follow-up time.
outcome hierarchy
Cardiovascular disease and major hemorrhage are secondary endpoints; original primary disability-free survival is C05.
population
Same 19114 ASPREE participants; primary prevention, mostly aged >=70. Species: Homo sapiens.
An unexpected mortality signal needs careful interpretation
ACCEPT_AS_CONTEXT
What was tested? This report examined death from any cause as a secondary outcome in the same aspirin-versus-placebo ASPREE trial. Analyses of causes of death were exploratory.
Who participated? The same 19,114 older participants were followed for a median of 4.7 years. This is a companion report from the original trial.
Findings Death rates were 12.7 with aspirin and 11.1 with placebo per 1,000 person-years. Cancer-related deaths contributed to the imbalance.
What was not demonstrated? This unexpected observation does not establish universal cancer causation or show that stopping aspirin reverses a cancer risk.
Harms and limitations The mortality imbalance is important, but it was a secondary finding. Looking separately at causes of death after the fact and testing multiple outcomes limit broad conclusions about cancer.
ASPREE observed a higher secondary all-cause mortality rate in the aspirin group, with cancer-related deaths contributing to the imbalance; this unexpected observation should not be presented as established universal cancer causation.
Exact accepted support, hierarchy and limitations
exact accepted proposition
ASPREE observed a higher secondary all-cause mortality rate in the aspirin group, with cancer-related deaths contributing to the imbalance; this unexpected observation should not be presented as established universal cancer causation.
exact numeric support
[{"absolute_difference_method": "LN arithmetic: 12.7 minus 11.1", "absolute_difference_unit": "events per 1000 person-years", "absolute_rate_difference": 1.6, "confidence_interval": [1.01, 1.29], "confidence_level": 0.95, "denominator_comparator": 9589, "denominator_intervention": 9525, "effect_scale": "hazard_ratio", "endpoint": "All-cause death", "estimate": 1.14, "rate_comparator": 11.1, "rate_intervention": 12.7, "rate_unit": "per 1000 person-years"}, {"confidence_interval": [1.1, 1.56], "confidence_level": 0.95, "denominator_comparator": 9589, "denominator_intervention": 9525, "effect_scale": "hazard_ratio", "endpoint": "Cancer-related death", "estimate": 1.31, "reported_percent_comparator": 2.3, "reported_percent_intervention": 3.1}]exact original limitations
Unexpected signal, secondary/exploratory hierarchy and multiple outcomes limit a broad causal cancer headline. Same participants as C05/C06, not replication. No inference that aspirin causes cancer in every age group or that stopping reverses a cancer risk. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.
exposure and comparator
Aspirin 100 mg daily Comparator: Placebo, original randomization
follow up
Median 4.7 years
interpretation
Hazard ratios compare event hazards during follow-up; they are not individual probabilities or cumulative risk differences. Confidence intervals express uncertainty. Person-years add participants’ follow-up time.
outcome hierarchy
All-cause mortality secondary; cause-specific analyses exploratory/post hoc; disability-free survival remains the parent primary endpoint.
population
Same 19114 ASPREE older adults; 1052 total deaths across arms during trial follow-up. Species: Homo sapiens.
Longer follow-up did not establish a delayed cardiovascular benefit
ACCEPT_AS_CONTEXT
What was tested? Researchers followed people after the ASPREE trial and compared major cardiovascular outcomes by original aspirin assignment. MACE is the combined major cardiovascular outcome used in this report.
Who participated? The original trial included 19,114 participants. Post-trial follow-up included 15,668 consenting participants without a MACE during the trial. Median post-trial follow-up was 4.3 years, with data through 2022.
Findings Extended follow-up did not demonstrate a statistically significant reduction in overall MACE by original aspirin assignment. The whole-period and post-trial-only estimates describe different periods.
What was not demonstrated? Later medication use was not continuously randomized. These results do not establish a causal effect of stopping aspirin or a rebound after stopping.
Harms and limitations Whole-period major bleeding remained an important adverse finding. Selection into follow-up and changes in treatment exposure limit interpretation of the post-trial estimates.
ASPREE extended follow-up did not demonstrate a statistically significant reduction in overall MACE by original aspirin assignment; post-trial estimates and whole-period hemorrhage must be kept separate from continuously randomized treatment effects.
Exact accepted support, hierarchy and limitations
exact accepted proposition
ASPREE extended follow-up did not demonstrate a statistically significant reduction in overall MACE by original aspirin assignment; post-trial estimates and whole-period hemorrhage must be kept separate from continuously randomized treatment effects.
exact numeric support
[{"confidence_interval": [0.94, 1.15], "confidence_level": 0.95, "denominator_note": "Original randomized cohort; post-trial consent/selection n=15668; endpoint-specific risk sets require full report.", "effect_scale": "hazard_ratio", "endpoint": "MACE, combined trial and post-trial period", "estimate": 1.04, "population_denominator": 19114}, {"confidence_interval": [1.01, 1.36], "confidence_level": 0.95, "effect_scale": "hazard_ratio", "endpoint": "MACE, post-trial period", "estimate": 1.17, "population_denominator": 15668}, {"confidence_interval": [1.1, 1.39], "confidence_level": 0.95, "effect_scale": "hazard_ratio", "endpoint": "Major hemorrhage, entire period", "estimate": 1.24, "population_denominator": 19114}]exact original limitations
Consent/survivor selection, in-trial-event exclusions and changing exposure. MACE differs from the original broader cardiovascular endpoint; do not pool HRs blindly. Abstract does not supply per-arm event counts or all absolute risks; none are invented. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.
exposure and comparator
Original aspirin assignment during trial; post-trial medication use not randomized Comparator: Original placebo assignment, with post-trial selection and treatment changes
follow up
Trial 2010-17; post-trial follow-up to 2022; median post-trial 4.3 years
interpretation
Hazard ratios compare event hazards during follow-up; they are not individual probabilities or cumulative risk differences. Confidence intervals express uncertainty. Person-years add participants’ follow-up time.
outcome hierarchy
Secondary/post-trial analysis of MACE and major hemorrhage; not a new primary-outcome randomization.
population
19114 originally randomized; 15668 without in-trial MACE consented to post-trial follow-up. Species: Homo sapiens.
Cancer incidence and cancer mortality tell different stories
ACCEPT_AS_CONTEXT
What was tested? Researchers examined new cancer diagnoses and cancer deaths during extended ASPREE/ASPREE-XT follow-up, comparing original aspirin assignments.
Who participated? Overall follow-up covered 19,114 original participants for a median of 8.6 years, using data through 2022. The post-trial-only analysis used 14,907 cancer-free participants who consented to follow-up.
Findings Original aspirin assignment did not show a significant reduction in cancer incidence. Cancer mortality was higher overall; the post-trial-only mortality estimate remained uncertain.
What was not demonstrated? These different periods and selected populations do not establish an effect of continuous long-term aspirin use or a causal effect of stopping it.
Harms and limitations Later treatment changed, and the post-trial participants were selected from cancer-free survivors. Those differences limit comparisons across the two time periods.
Over median 8.6-year ASPREE/ASPREE-XT follow-up, original low-dose aspirin assignment did not show a significant reduction in cancer incidence; higher overall cancer mortality and the post-trial-only uncertain mortality estimate refer to different periods and selected populations.
Exact accepted support, hierarchy and limitations
exact accepted proposition
Over median 8.6-year ASPREE/ASPREE-XT follow-up, original low-dose aspirin assignment did not show a significant reduction in cancer incidence; higher overall cancer mortality and the post-trial-only uncertain mortality estimate refer to different periods and selected populations.
exact numeric support
[{"confidence_interval": [0.92, 1.05], "confidence_level": 0.95, "effect_scale": "hazard_ratio", "endpoint": "Cancer incidence, whole period", "estimate": 0.98, "population_denominator": 19114}, {"confidence_interval": [1.03, 1.29], "confidence_level": 0.95, "effect_scale": "hazard_ratio", "endpoint": "Cancer-related mortality, whole period", "estimate": 1.15, "population_denominator": 19114}, {"confidence_interval": [0.82, 1.01], "confidence_level": 0.95, "effect_scale": "hazard_ratio", "endpoint": "Cancer incidence, post-trial only", "estimate": 0.91, "population_denominator": 14907}, {"confidence_interval": [0.83, 1.25], "confidence_level": 0.95, "effect_scale": "hazard_ratio", "endpoint": "Cancer mortality, post-trial only", "estimate": 1.02, "population_denominator": 14907}]exact original limitations
Post-trial treatment changes and cancer-free survivor/consent selection; time windows and denominators differ. Per-arm absolute cancer risks are not qualified from the abstract. PMCID is a locator; index release date is 2027, so current free full-text availability is not assumed. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.
exposure and comparator
Original daily aspirin 100 mg assignment until study-drug cessation Comparator: Original placebo assignment; later exposure observational
follow up
Reported outcome data 2010-2022; median 8.6 years, with separate post-trial analysis
interpretation
Hazard ratios compare event hazards during follow-up; they are not individual probabilities or cumulative risk differences. Confidence intervals express uncertainty. Person-years add participants’ follow-up time.
outcome hierarchy
Cancer incidence/mortality analyses within extended follow-up, not parent primary disability-free survival or a new randomization.
population
19114 original participants, 3448 incident cancers and 1173 cancer deaths overall; 14907 cancer-free consenting extension participants for legacy analyses. Species: Homo sapiens.
Summary
ASPREE compared enteric-coated aspirin 100 mg daily with placebo in 19,114 independent older adults in Australia and the US, mostly aged 70 or older (specified US minority groups from age 65), without overt cardiovascular disease, dementia or disability limiting independence. Median trial follow-up was 4.7 years.The main combined outcome rate was 21.5 with aspirin and 21.2 with placebo per 1,000 person-years. Aspirin did not significantly improve this outcome during the trial. The selected, independent population and finite follow-up limit broader conclusions. Taking the assigned tablets declined over time. Major bleeding is an important separate result in the same trial, shown alongside this finding.Cardiovascular event rates were 10.7 with aspirin and 11.3 with placebo per 1,000 person-years, without a statistically clear reduction. Major bleeding rates were 8.6 and 6.2 per 1,000 person-years, respectively. Major bleeding was more frequent with aspirin. The reported rate difference is 2.4 additional major bleeds per 1,000 person-years. This rate difference is not the percentage of people who bleed over the whole trial.Death rates were 12.7 with aspirin and 11.1 with placebo per 1,000 person-years. Cancer-related deaths contributed to the imbalance. The mortality imbalance is important, but it was a secondary finding. Looking separately at causes of death after the fact and testing multiple outcomes limit broad conclusions about cancer.Extended follow-up did not demonstrate a statistically significant reduction in overall MACE by original aspirin assignment. The whole-period and post-trial-only estimates describe different periods. Whole-period major bleeding remained an important adverse finding. Selection into follow-up and changes in treatment exposure limit interpretation of the post-trial estimates.Original aspirin assignment did not show a significant reduction in cancer incidence. Cancer mortality was higher overall; the post-trial-only mortality estimate remained uncertain. Later treatment changed, and the post-trial participants were selected from cancer-free survivors. Those differences limit comparisons across the two time periods.ASPREE-XT is follow-up of this parent trial, not another randomized trial. Later medication exposure was observational. The original primary outcome, secondary outcomes, whole-period results and post-trial-only results must be read separately.
Intelligence Take
Makes the trial’s main outcome visible before secondary findings or claims about longer independent life. Shows the uncertain cardiovascular result together with the important bleeding harm. Keeps an important adverse finding visible without turning it into a general cancer-causation claim. Separates the original treatment assignment from what people took later and keeps analysis periods distinct. Helps readers distinguish new diagnoses from deaths and whole-period findings from selected post-trial results.
Limitations
Healthy independent baseline and finite follow-up limit extrapolation to frailty, established CVD or lifelong use. Adherence declined; trial assignment is not continuous exposure. Composite null is not equivalence, component-specific null or a direct stopping-trial result. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.No individual benefit/bleeding prediction or personalized starting/stopping recommendation. Secondary prevention and other doses/populations are not tested here. Source-qualified rate differences do not supply cumulative risk, NNT/NNH, equal severity or long-term lifetime benefit. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.Unexpected signal, secondary/exploratory hierarchy and multiple outcomes limit a broad causal cancer headline. Same participants as C05/C06, not replication. No inference that aspirin causes cancer in every age group or that stopping reverses a cancer risk. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.Consent/survivor selection, in-trial-event exclusions and changing exposure. MACE differs from the original broader cardiovascular endpoint; do not pool HRs blindly. Abstract does not supply per-arm event counts or all absolute risks; none are invented. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.Post-trial treatment changes and cancer-free survivor/consent selection; time windows and denominators differ. Per-arm absolute cancer risks are not qualified from the abstract. PMCID is a locator; index release date is 2027, so current free full-text availability is not assumed. Access qualification: Official indexed author abstract only; full methods and supplements unreviewed.