How to translate clinical trial results into gain in healthy life expectancy for individual patients
BMJ 2016; 352 doi: https://doi.org/10.1136/bmj.i1548 (Published 30 March 2016) Cite this as: BMJ 2016;352:i1548
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
This research methods and reporting article is indeed aimed at illustrating how methods of lifetime treatment effect prediction work. Aspirin in healthy women was chosen as an illustrative example. Our findings that early prevention results in larger gain in healthy life-expectancy, however, can also be generalized to other types of interventions. We agree that these include not only pharmacological interventions, but also life-style interventions, for example smoking cessation (BMJ. 2004 Jun 26;328(7455):1519). In any case, lifestyle interventions are the first line of treatment. Whether additional pharmacological interventions at a young age are beneficial depends on individual risk factors. Of course, individual treatment effect predictions must be considered as estimates, not as prophecy. However, as shown by temporal validation, these estimates have proven adequate validity, even after long-term follow-up. Medical decision-making, as always, should be based on shared consideration of anticipated benefits and harms.
Competing interests: No competing interests
This article [1] comes after many other articles based on models and trials such as the Systolic Blood Presssure Intervention Trial (SPRINT) [2], Heart Outcomes Prevention Evaluation (HOPE-3) [3] and this study modeling primary drug prevention [4].
These four studies, taken as examples, advocate the use of multidrug prevention in younger and healthier asymptomatic adults with intermediate risk for cardiovascular diseases. All these studies have common features -- i.e., they draw general conclusions from peculiar approaches or highly selected populations.
This was the case for the SPRINT trial. The SPRINT trial was a randomized controlled unblinded trail that compared two arms of a sample of 9361 patients with a systolic blood pressure higher than 130 and increased cardiovascular risk but without diabetes. The target SBP for one group was 120 mmHg and 140 mm Hg for the other. The eligibility criteria were so selective that, in a country where nearly the third of the population has high blood pressure, it needed two years and amendments to these criteria to complete the enrollment of 9361 patients in 105 centres.
The HOPE-3 trial is a randomized controlled trial funded by Astra Zeneca that compares the use of rosuvastatin versus placebo in a an intermediate cardiovascular risk patient group. The top sale drug of this pharmaceutical company is, precisely, rosuvastatin [5]. S.Yusuf, the main author of the study is also a strong supporter of the polypill concept: “The idea behind the polypill originally was to give it to everybody in the population over age 55 years”[6] [7].
The four studies don’t reveal major new facts. And the findings of HOPE-3 are nearly the same as other studies about statins [8]. The median follow-up in the SPRINT trial was only 3.2 years because it was stopped earlier than anticipated owing to “positive outcomes” in the composite cardiovascular criteria and in cardiovascular deaths and all cause mortality. Some called then for the revision of blood pressure guidelines, and Paul Whelton, chair of the trial, concluded that “the lower, the better”.
In the current study, the favorable results can be entirely explained by the mechanical action of the analysis tools that were chosen, which are lifespan and competing risks. Because lifespan is mechanically reduced in older patients and other cause mortality also increases with age. So the two criteria favor younger patients for positive outcomes.
The main novelty allowing the drawing of such radical conclusions as advocating treatment of healthy patients then lies in the interpretation of facts and the particular approach endorsed by authors. So, considering the major safety and economic concerns that arise with the systematic multidrug treatment of asymptomatic healthy young adults, we must wonder if these studies are legitimate to recommend such large changes which can amount to a radical change in society. Because if accepted, this conclusion may lead to dropping any attempt at lifestyle prevention and medicalizing systematically healthy patients at intermediate or low cardiovascular risk.
Actually, all these four studies have at least two major flaws.
The first flaw is that they don’t take into account, minimize or deny, the risk of severe and impairing adverse effects of a multidrug treatment. For example, in the SPRINT trial there were 870 severe adverse effects requiring admission to the emergency room in the standard treatment group as compared to 1140 in the intensive treatment group, that is, 32% more. Oddly enough, there was less dizziness, a current adverse effect of high blood pressure treatment, in the intensive treatment group in which 24.3% people were treated with 4 or more drugs, as compared with 6.9 % in the standard treatment group. This was at the cost of a monthly follow-up for the intensive treatment group only.
In the current study from Dorresteijn side effects are outlined in very vague terms, although we know adverse reactions would be multiplied by multiple treatments, because of the individual potential of every drug for adverse reactions but also because of the risk for multiple drug interactions. Moreover, the increase in adverse events would be proportional to the additional number of patients treated and the additional duration of treatment for these patients.
A French prospective study has shown that 8,37% admission to the emergency room for polymedicated patients 65 years and over was related to adverse drug events [9].
The second flaw, in this particular study, is pretending that individual predictions are possible whereas they are not.
There are three reasons why individual predictions are not possible.
The first is that calculation of cardiovascular risk with a limited number of variables doesn’t define individual risk but a group’s risk which size depends on the size of the sample considered.
The second one is that modifiable risk factors, mainly lifestyle factors determined by environnement, by choices and behavior were not taken at all into account. However, even a modest change in only one lifestyle factor for an individual, i.e. stop smoking, increasing exercise or a healthier diet could push this individual out of that group at risk.
The third one is that there are huge variations in incremental relative risk for individual patients with variable variations [10]. That means that risk calculators are not reliable in predicting individual risks.
In conclusion, high blood pressure, high cholesterol, diabetes and smoking are not fatalities. They are the result of an unfavorable environment that can be controlled by regulation policies which are highly cost-effective [11] and have few adverse effects. They are also the consequence of bad choices and unsuitable behavior like being sedentary. The benefits of the correction of lifestyle risk factors far exceed cardiovascular benefits.
The biological risk factors related to degenerative diseases are also increasing in emerging countries. Let them settle and provide drug treatments is certainly the most costly and the worst response. All these factors can be modified by better education, information, better living conditions and a better environment.
[1] http://www.bmj.com/content/352/bmj.i1548?etoc=
[2] http://www.nejm.org/doi/full/10.1056/NEJMoa1511939#t=article
[3] http://www.nejm.org/doi/10.1056/NEJMoa1600176
[4] http://openheart.bmj.com/content/3/1/e000343.full.pdf+html
[5] http://www.statista.com/statistics/267807/astrazeneca-top-products-based...
[6] http://www.tctmd.com/show.aspx?id=134624
[7] http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3102053/
[8] http://www.tctmd.com/show.aspx?id=134624
[9] http://www.ncbi.nlm.nih.gov/pubmed/19591522
[10] http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4561470/
[11] http://www.bmj.com/content/352/bmj.i33
Competing interests: No competing interests
Re: How to translate clinical trial results into gain in healthy life expectancy for individual patients
Translating clinical trial results into measurable quantities that explain the importance of health gain in individual patients. The article explains in simple terms about the competing risks and competing risk adjusted lifetime prediction models. Competing risk event is the event other than the one of interest which alters the probability of experiencing the event of interest.1 Managing competing risks during the analysis of data is essential to avoid misinterpretations especially the effects on the risk estimates. Standard survival model determine the risk of the event having occurred by a certain time, an important assumption is that censoring is not associated with an altered chance of the event occurring at any given moment.2 The patient who has not experienced the event at the end of follow-up is said to be censored.
The hazard ratios for Cox proportional hazards regression models for each event are calculated for comparison of the cumulative incidence functions such as one daily rate ratio for infection (the event of interest) and one daily rate ratio for discharge without infection (the competing event). Fine and Gray model is used to make analysis of sub-distribution hazard ratios which is a time- averaged risk comparison, which extends to overall risk ratios.2These approaches help understand how risk factors are directly and indirectly associated with events of interest when they are competing events.
Kaplan- Meier method is used to obtain the nonparametric estimate of the cumulative incidence when the data consist of patients who experience an event and censored individuals. Censoring mechanism is assumed to be non-informative in this approach. The cumulative incidence function should be calculated by taking into account the competing risks.1
Number needed to treat (NNT), typically used for treatment with medication, is an epidemiological measure of a health care intervention translates its effectiveness. It’s the average number of patients who need to be treated to prevent one additional bad outcome. 3 In a clinical trial, the number of patients that need to be treated to for one to benefit compared to a control. In an ideal situation NNT is 1, where everyone benefits from treatment and no one improves with control. There is an inverse relationship between NNT and how effective is the treatment. 4
For illustrating the survival gain in individuals based on RCT results, authors describe 100 mg aspirin RCT conducted on 27,939 women in a primary prevention study of cardiovascular disease.5 They preferred the analysis using competing risk adjusted lifetime model as compared to NNT as the latter has shown generally lower results in patients at high risk. The hazard ratio favours aspirin but not significant as p is 0.91 (95% CI is 0.80 to 1.03).5 By using the cumulative survival, the aspirin treatment effect, in terms of life gain for a particular 48 years old patient was 0.6 years or 7 months.5 Authors further validated the results of the model with Kaplan- Meier survival estimates.
This approach as illustrated can be adopted for the disease such as beta thalassemia. In beta thalassemia major patients due to repeated transfusions there is increased iron absorption which leads to iron overload6. These patients may present with reduced exercise tolerance or right- sided heart failure features due to cardiac iron overload. Desferroxamine, a chelating drug which has led to signficant life expectancy gains in thalassemia major patients. In chronic iron overload it is an important first-line treatment. The drug is given at night, subcutaneously over 8 to 12 hours, 3 to 7 times, a week using a pump. As the infusion regime is demanding, therefore the compliance can be difficult7. In some cases intensive chelation can reduce cardiac iron overload and even resolution of heart failure6.
References
1. Satagopan JM, Ben-Porat L, Berwick M et al. A note on competing risks in survival data analysis. Br J. Cancer.2004; 91: 1229–1235.
2. Wolkewitz M, Cooper BS, Bonten MJ et al. Interpreting and comparing risks in the presence of competing events. MJ 2014; 349:g5060.
3. Laupacis A, Sackett DL, Roberts RS. An assessment of clinically useful measures of the consequences of treatment. N. Engl. J. Med 1988; 318 (26): 1728–33.
4. Number Needed to Treat. Bandolier. Retrieved 2016-04-12.
5. Dorresteijn JA, Kaasenbrood L, Cook N et al. How to translate clinical trial results into gain in healthy life expectancy for individual patients. BMJ 2016; 352: i1548.
6. Aessopos A , Berdoukas V , Tsironi M, The heart in transfusion dependent homozygous thalassaemia today--prediction, prevention and management. Eur J Haematol. 2008;80(2):93-106. Epub 2007.
7. Roberts DJ, Rees D, Howard J, et al; Desferrioxamine mesy late for managing transfusional iron overload in people with transfusion-dependent thalassaemia. Cochrane Database Syst Rev 2005,(4): CD004450.
Competing interests: No competing interests