| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
First Published Online December 10, 2009
The Oncologist, Vol. 14, No. 12, 1169-1174, December 2009; doi:10.1634/theoncologist.2009-0206
© 2009 AlphaMed Press
OPEN ACCESS ARTICLE
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Commentaries |
Putting Insulin Glargine and Malignancies into Perspective
Medizinische Klinik und Poliklinik I, Universitätsklinikum Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany
Key Words. Insulin • Malignancy • Glargine • Colon • Pancreas • Breast
Correspondence: Gerhard Ehninger, M.D., Medizinische Klinik und Poliklinik I, Universitätsklinikum Carl Gustav Carus, Technische Universität Dresden, Fetscherstrasse 74, 01307 Dresden, Germany. Telephone: 49-351-4584190; Fax: 49-351-4585362; e-mail: gerhard.ehninger{at}uniklinikum-dresden.de, Web site:http://www.mk1dd.de
Received August 31, 2009; accepted for publication October 31, 2009; first published online in THE ONCOLOGIST Express on December 10, 2009.
Disclosures: Gerhard Ehninger: None; Alexander H. Schmidt: None.
The content of this article has been reviewed by independent peer reviewers to ensure that it is balanced, objective, and free from commercial bias. No financial relationships relevant to the content of this article have been disclosed by the authors or independent peer reviewers.
 |
INTRODUCTION |
|---|
Top
Introduction
Observational Data" ... while society expects due diligence in the detection of serious drug side-effects, claims of harm not backed by adequate evidence can provoke unnecessary alarms and anxieties, and seriously interfere with good medical practice." Pocock and Smeeth [1].
Patients with type 2 diabetes have been shown to have a higher risk for cancer. The excess risk compared with nondiabetic matched individuals is 30% for colon cancer, 50% for pancreatic cancer, and 20% for postmenopausal breast cancer [2–4]. These are, however, also associated with obesity, making the causal relationship complex. Further to this, pharmacotherapy for the treatment of diabetes has been shown to be associated with greater cancer mortality, with metformin being neutral and sulfonylureas (hazard ratio [HR], 1.3; 95% CI, 1.1–1.6) and insulin (HR, 1.9; 95% CI, 1.5–2.4) showing higher mortality [5].
Insulin glargine is a long-acting insulin analog that is identical to human insulin except at position 21 of the A-chain (Asp 
Gly) and two additional terminal Arg in the B-chain. It was developed in response to a clinical demand for prolonged stable action and, most of all, a lower risk for hypoglycemia. These advanced properties have led to its widespread use as basal insulin in clinical practice. It is, however, unknown whether or not insulin glargine and human insulin are different with respect to mitogenicity. They have so far only been compared in in vitro studies showing that insulin glargine had greater receptor affinity and mitogenic potency [6], but in vivo metabolism of insulin glargine at the injection site and in the circulation suggest low mitogenic potency [7].
|
OBSERVATIONAL DATA |
|---|
Recently, data from an observational, retrospective cohort study in Germany were published. A significantly lower incidence of cancer for patients on insulin glargine in the unadjusted analysis as well as after adjustment for age and sex (HR, 0.86; 95% CI, 0.79–0.94) compared with patients on human insulin [8] were shown. This result was reversed by an additional step, adjusting for insulin doses (HR, 1.14; 95% CI, 1.05–1.24). The results of this study were submitted to Diabetologia, the editors of which realized a number of limitations [9]. "These ranged from biological implausibility, given the short median period (1.31 years for insulin glargine) of exposure to each of the insulins, to the limited overlap between the dose ranges, the unexplained effect of insulin glargine on all-cause mortality, the lack of overall difference in cancer risk among the four insulins in the crude analysis, failure to correct for BMI in the dose–response analysis, and a number of technical considerations." Three of the six reviewers recommended rejecting the article. Publication of this "hypothesis-generating" data was finally made, conditional upon other retrospective analyses, which were initiated subsequently [9] (Table 1).
|
A Swedish study identified patients receiving prescriptions for insulin in 2005 from different registries [10] and linked their records to cancer registry data for the two subsequent years. No significant difference in cancer rates was found between patients on insulins other than insulin glargine and those on insulin glargine plus other insulins. Patients who were treated with insulin glargine alone, however, had a higher risk for breast cancer than those on other insulins (relative risk, [RR], 1.99; 95% CI, 1.31–3.03). Risks for other malignancies did not differ among the various groups. The authors stated that they had a priori expected a higher risk for all or none of the analyzed cancer types and concluded: "We found no statistically significant results for prostate or gastrointestinal cancer, which, again, strengthens the interpretation that the breast cancer results were due to random fluctuation" [10].
A Scottish analysis [11] was based on a national diabetes database. Those investigators analyzed all patients exposed to insulin therapy for the calendar years 2002–2004 and matched them with cancer registry data up to the end of 2005. There was a significantly lower risk for all cancer types with no greater overall risk for breast cancer with insulin glargine. The authors stated that they added a further analysis to examine if the analysis of Hemkens et al. [8] might be biased, because it "was restricted to users of insulin glargine who were not using any other insulins concurrently." For this subgroup, they found a higher overall cancer rate (HR, 1.55; 95% CI, 1.01–2.37), with more cases of breast cancer in those on insulin glargine than those on nonglargine insulins (HR, 3.39; 95% CI, 1.46–7.85; p = .004). The authors underlined that they do not expect this to be caused by a different insulin effect but by a biased estimate: "We wanted to understand whether restricting the analysis to such a subgroup could lead to a biased estimate of any overall relationship between cancer and insulin glargine, not because we were hypothesising differential effects of insulin glargine that are dependent on concomitant insulin."
The authors [11] furthermore discussed that patients using glargine only constituted only a small subset with an "allocation bias" toward older and "generally less healthy" patients. They concluded that: " ... our primary hypothesis that exposure to insulin glargine was associated with an adverse effect on total cancer rates or site-specific cancer was refuted, and these data are reassuring in this regard" [11].
The U.K. general practice analysis [12] was based on The Health Improvement Network database, representing the most detailed database with carefully defined subcategories according to diabetes therapy. This analysis showed no significant differences in malignancy risk for the four insulin-treated groups (insulin glargine only, neutral protamin Hagedorn [NPH] insulin only, human biphasic insulin only, and analog biphasic insulin only).
All publications lacked prespecifications or defined objectives, which causes multiplicity, a serious issue that cannot be controlled for. According to the editorial, "the overall null hypothesis was that people treated with insulin analogues were not more likely to be diagnosed with cancer" [9]. With regard to this primary objective, none of the additional studies confirmed the findings of the Hemkens et al. [8] analysis.
|
CRITICAL APPRAISAL OF HEMKENS ET AL. [8] |
|---|
The Hemkens et al. [8] paper was recently commented on by Pocock and Smeeth [1], in an article unambiguously entitled "Insulin glargine and malignancy: An unwarranted alarm." The comment was recently published in The Lancet, concluding with respect to the Hemkens et al. [8] analyses that " ... the methods used are fundamentally flawed, making the conclusions unsupportable" and highlighting "two serious errors, which make the article's findings uninterpretable." These errors are the a posteriori allocation of patients into treatment groups based on follow-up information and the use of mean insulin doses as observed during the follow-up period as baseline covariates in a survival analysis. They further underlined that the emphasized findings for breast cancer in the Swedish study, which was not confirmed in the Scottish analyses, should not be overestimated because of a lack of prespecification and multiplicity.
|
INCLUSION CRITERIA CREATE DISTINCTIVE OBSERVATIONAL GROUPS |
|---|
The Hemkens et al. [8] study suffers from important imbalances between the observational groups (Table 2) that cannot be controlled by adjustment, either because they are not available from the data set, for example, body mass index (BMI) or smoking habits, or because adjustment of actual medically distinctive groups can be misleading. The latter include characteristics of patients in the human insulin group that are excluded in the glargine group: use of short-acting insulins, type 1 diabetes, and use of other (supplementary insulin therapy, conventional therapy) and intensified (intensified conventional therapy) regimens. The author group decided to solely compare patients receiving a long-acting insulin glargine only with patients receiving human insulin irrespective of it being a long- or short-acting insulin or even a combination of both. This is unwise, creating two biologically different cohorts. Patients sufficiently treated with a long-acting or "basal" insulin only still produce enough endogenous insulin to control postprandial blood glucose. On the other hand, patients receiving both short- and long-acting insulin are mostly lacking endogenous insulin and may even include patients with type 1 diabetes. According to the inclusion criteria, there would be patients with basal insulin only in the human insulin groups as well, but the report gives no deeper information on this. It states, however, representativeness for Germany, which means that the human insulin group would be dominated by patients who need a substitution with basal and bolus insulins because of a severe lack of endogenous insulin. This deduction is empirically supported by the insulin doses for the observational groups [8]—22 U/day in the insulin glargine group versus 37 U/day in the human insulin group; 95% quartiles differ by 59 U/day versus 100 U/day, respectively.
|
All these imbalances are mainly a consequence of the chosen inclusion criteria and could have been prevented easily either by comparing long-acting glargine with long-acting human insulin exclusively or by comparing insulin glargine plus optional bolus insulins with human basal insulin plus optional bolus insulins. However, the two fundamental flaws of the Hemkens et al. [8] study as revealed by Pocock and Smeeth [1] would still persist as a result of the very nature of the available data.
|
HIGHER CALCULATED CANCER INCIDENCE IS CAUSED BY ADJUSTMENT FOR DOSE |
|---|
It must be considered that the Hemkens et al. [8] study primarily shows a significantly lower cancer incidence for patients treated with insulin glargine than for those treated with human insulin. This effect persists after adjustment for age and gender, with an HR of 0.86 (95% CI, 0.79–0.94). A further adjustment for dose, based on the observed differences (22 U/day with insulin glargine versus 37 U/day with human insulin) reverses this relationship (HR, 1.14; 95% CI, 1.05–1.24). This adjustment was heavily criticized in The Lancet article by Pocock and Smeeth [1] to be based on "serious errors [which] make the article's findings unsupportable."
Beyond that, one should consider that the substantial difference in dose that is subject to this flawed adjustment can partly or totally be explained as an artificial effect resulting from selective inclusion criteria rather than medical reality. As described above, the inclusion criteria result in a group of patients with a higher degree of residual endogenous insulin in the glargine group, enabling those patients to be treated without bolus insulins, being compared with human insulin patients, many of whom are treated with a combination of basal and bolus insulins, indicating low levels of endogenous insulin secretion. For this reason, an adjustment for mean external insulin doses as performed by Hemkens et al. [8] causes a systematic error toward higher cancer risks for insulin glargine because it does not consider the different levels of endogenous insulin.
|
IS THERE A RELATIONSHIP BETWEEN DOSE AND CANCER INCIDENCE? |
|---|
Hemkens et al. [8] present HRs of glargine versus human insulin for three classes of patients, with mean daily doses for each insulin type of <20 U/day, 20–40 U/day, and >40 U/day. They obtained a significantly higher HR for the glargine group for only the >40 U/day dose, of 1.59 (95% CI, 1.30–1.94). The HR for the 20–40 U/day group (1.01; 95% CI, 0.88–1.16) was even lower than that for <20 U/day group (1.13; 95% CI, 0.98–1.30).
This result raises questions regarding the validity of the regression model used by Hemkens et al. [8] because this model delivered a nearly linear increase in HR with dose from 1.09 (10 U/day) over 1.19 (30 U/day) to 1.31 (50 U/day). Because the basis of the regression calculations is fundamentally flawed anyway, this additional problem seems, however, to be of only minor relevance.
There was substantial imbalance within the high-dose class of >40 U/day. Only 13.5% of the insulin glargine patients but 46% of the human insulin patients constituted this high-dose class. This again is caused by the study design. Patients with an overall insulin use >40 U/day (human insulin group, both short- and long-acting insulins) were compared with patients with >40 U/day of basal insulin alone (glargine group), ignoring that these patients additionally had a relevant amount of endogenous insulin, sufficient to avoid the necessity for postprandial bolus insulin.
Differences in the dose used may be explained by the BMI of the patients. On the other hand, a rise in BMI is positively correlated with cancer [13]. A recent meta-analysis made from 31 studies showed that an increase in BMI of 5 kg/m2 led to a significant increase in cancer rates for, among others, colon cancer, pancreatic cancer, and postmenopausal breast cancer in women [14]. The BMI is, however, not available from the AOK (Allgemeine Ortskrankenkasse, the main statutory health insurance in Germany) data set. Therefore, the missing adjustment for BMI makes it impossible to state that there is a dose dependency of cancer risk at all. The Swedish study tried to confirm this dose relationship but failed [10]. Because the Swedish data were adjusted for BMI, this supports the notion that the dose relationship described by Hemkens et al. [8] is not real but a confounding effect of BMI.
|
ARE THE ANALYSES SUPPORTED BY AVAILABLE RANDOMIZED CONTROLLED TRIAL EVIDENCE? |
|---|
Erbel et al. [15] reported data on 125 patients with diabetes after partial or total pancreatectomy resulting from pancreatic carcinoma followed up for a median duration of 22 months. The median survival time was 15 months, with no difference in the time-dependent proportion of patients who survived between insulin glargine users and those with nonglargine insulin treatment (p = .04). The authors concluded from their analysis that regular human insulin and insulin glargine may be used to treat diabetes in patients with pancreatic cancer.
Rosenstock et al. [16] analyzed the data set from a randomized long-term study (n = 1,024) that compared insulin glargine with human NPH insulin for 5 years in the management of type 2 diabetes. The mean cumulative exposure in that study was >4 years. The overall numbers of patients with neoplasms (using the standard Medical Dictionary for Regulatory Activities coding dictionary) occurring during the trial were similar in the two treatment groups—57 patients (11.1%) in the insulin glargine group versus 62 patients (12.3%) in the NPH insulin group, with an RR for insulin glargine of 0.90 (95% CI, 0.64–1.26). The RR estimate for all malignant breast tumor cases was calculated to be 0.59, with a 95% CI of 0.14–2.44.
Of major importance is an ongoing, placebo-controlled outcome trial testing the injection of insulin glargine in patients with dysglycemia to determine whether targeting normal fasting glucose levels with insulin glargine reduces cardiovascular outcomes compared with standard management. Participants have been followed up for an average of 4 years to date for a total exposure of >50,000 person years. Data are collected related to all major serious health outcomes, including cancers. In light of the questions raised by the recent publications, this committee of experts recently reviewed data related to cancers in both treatment groups and concluded that there is no cause for concern and no reason to alter the design of the study for safety reasons [17].
|
HEMKENS ET AL. [8] WITHIN THE GERMAN CONTEXT |
|---|
The Hemkens et al. [8] study was published by a joint author team of the German Institute for Quality and Efficiency in Health Care (IQWiG) and the Research Institute of the AOK. Of note might also be that no ethical committee was involved and the data were not anonymized (the pseudonymization process remained unspecified) despite the fact that patients did not give their informed consent for the purpose of the study.
There has been widespread attention to these publications and the analysis by Hemkens et al. [8] in particular, with a perceived higher cancer risk for patients on insulin glargine. This awareness was further boosted in the mass media, including television statements by the head of the IQWiG and senior author of the Hemkens et al. [8] study arguing that, in Germany alone, an extrapolated 400–4,000 potential excess cancer cases per 1.5 years might be a result of the use of insulin glargine [18, 19]. Such statements are in noticeable contrast to the hypothesis-generating nature of the Hemkens et al. [8] study as explicitly defined in the publication. The publicly perceived higher risk for cancer led to substantial uncertainty among patients and physicians and caused withdrawals of insulin glargine because of cancer fears. One of the authors (G.E.) of the current comment previously opposed this perception and its consequences [20] because of substantial flaws in the study leading to medically wrong conclusions. We even would be willing to go further in claiming that the obvious deficiencies of these analyses were recognized by the authors but were dismissed.
The IQWiG was founded in 2004. One of its tasks is to evaluate the benefit of pharmaceutical drugs with the intention of informing the federal committee Gemeinsamer Bundesausschuss (G-BA). The G-BA is the heading board within the self-governing health care system and makes binding decisions on reimbursement of drugs.
A recent report of the IQWiG (March 2009) evaluated the benefit of long-acting insulins in the treatment of type 2 diabetes [21]. During the planning of this assessment, staff members of the IQWiG copublished a Cochrane review that documented a lower risk for symptomatic nocturnal hypoglycemia [22]. The IQWiG assessment itself describes a significantly lower risk for severe as well as symptomatic hypoglycemias with insulin glargine than with long-acting human insulin (NPH) as a result of a meta-analysis based on individual patient data [21]. In contrast to the benefit of a lower risk for severe hypoglycemia with insulin glargine, the superiority in terms of a lower risk for symptomatic hypoglycemia was questioned by the report later on because of a newly introduced "adjusted decision limit." This limit refers to a study by Wood et al. [23] indicating that, without blinding, the effects of interventions tend to be overestimated in trials with subjective outcomes. However, they concede that there is "no simple or widely accepted method to integrate assessments of risk of bias into meta-analyses." Therefore, the adjusted decision limit can be seen as an invention by the IQWiG. Despite these findings, the resumé documented no incremental benefit for insulin glargine over human insulin (NPH insulin).
A potential influence on cancer rates by insulin glargine was also addressed in that report [21], which included only randomized, controlled trials, in line with their general drug assessment policy. The report identifies only one randomized, long-term study that compares insulin glargine with NPH insulin [16]. Although the study showed no evidence for a greater cancer risk with insulin glargine, the report concluded that the number of patients (n = 1,024) might have been too small for a definitive judgment of cancer rates. The report did not refer to the analysis by Hemkens et al. [8], which had already been submitted at the time of hearing of the draft report and publication of the final report.
|
CONCLUSIONS |
|---|
Preclinical studies have pointed to possible greater mitogenicity of insulins. Nonetheless, insulin treatment is a cornerstone in the clinical armamentarium to prevent the fatal consequences of diabetes. Whether or not there are differences between different insulin preparations is an important question, but only methodologically sound analyses with clear prespecifications and scientific discussion can contribute to our knowledge. The present analyses and their perception in the media caused and still do cause a lot of concern, and the European Medicines Agency (EMEA) immediately started a review and disclosed, in a press release on July 23, 2009: "Concerns over a possible relationship between these medicines and cancer, in particular breast cancer, were raised by four recently published registry studies. The Committee carried out an in-depth review of these studies and their outcomes. Due to methodological limitations the studies were found to be inconclusive and did not allow a relationship between insulin glargine and cancer to be confirmed or excluded. In addition, the Committee noted that the results of the studies were not consistent." The conclusion of the EMEA was: "... available data does not provide a cause for concern and that changes to the prescribing advice are therefore not necessary" [24].
Nevertheless, concern remains among patients, their relatives, and even their physicians. Hypothesis-generating analyses should only be published if there is a balanced discussion. This was, however, not apparently intended by the authors because they distributed their findings aggressively to the lay press through the IQWiG with alarming headlines that read "Insulin analogue glargine possibly increases cancer risk."
The lack of information about major cancer risk factors (e.g., BMI and smoking habits), the comparison of two very different groups of patients, and the use of highly questionable statistical approaches make the paper by Hemkens et al. [8] an example of what happens when a comparison is made between "apples and oranges." Worse yet, when appropriate statistical methods are abused or abandoned, the result confuses rather than clarifies a field that needs sound science rather than sensationalism. The words of Peacock and Smeeth ring in one's ears: " ... claims of harm not backed by adequate evidence can provoke unnecessary alarms and anxieties, and seriously interfere with good medical practice" [1].
|
AUTHOR CONTRIBUTIONS |
|---|
Conception/Design: Gerhard Ehninger, Alexander Schmidt
Collection and/or assembly of data: Gerhard Ehninger, Alexander Schmidt
Data analysis and interpretation: Gerhard Ehninger, Alexander Schmidt
Manuscript writing: Gerhard Ehninger, Alexander Schmidt
Final approval of manuscript: Gerhard Ehninger, Alexander Schmidt
|
REFERENCES |
|---|
- Pocock SJ, Smeeth L. Insulin glargine and malignancy: An unwarranted alarm. Lancet 2009;374:511–513.[CrossRef][Medline]
- Huxley R, Ansary-Moghaddam A, Berrington de Gonzalez A et al. Type-II diabetes and pancreatic cancer: A meta-analysis of 36 studies. Br J Cancer 2005;92:2076–2083.[CrossRef][Medline]
- Larsson S, Mantzoros C, Wolk A. Diabetes mellitus and risk of breast cancer: A meta-analysis. Int J Cancer 2007;121:856–862.[CrossRef][Medline]
- Larsson SC, Orsini N, Wolk A. Diabetes mellitus and risk of colorectal cancer: A meta-analysis. J Natl Cancer Inst 2005;97:1679–1687.
[Abstract/Free Full Text] - Bowker SL, Majumdar SR, Veugelers P et al. Increased cancer-related mortality for patients with type 2 diabetes who use sulfonylureas or insulin. Diabetes Care 2006;29:254–258.
[Abstract/Free Full Text] - Kurtzhals P, Schaffer L, Sørensen A et al. Correlations of receptor binding and metabolic and mitogenic potencies of insulin analogs designed for clinical use. Diabetes 2000;49:999–1005.[Abstract]
- Agin A, Jeandidier N, Gasser F et al. Glargine blood biotransformation: In vitro appraisal with human insulin immunoassay. Diabetes Metab 2007;33:205–212.[CrossRef][Medline]
- Hemkens LG, Grouven U, Bender R et al. Risk of malignancies in patients with diabetes treated with human insulin or insulin analogues: A cohort study. Diabetologia 2009;52:1732–1744.[CrossRef][Medline]
- Smith U, Gale EA. Does diabetes therapy influence the risk of cancer? Diabetologia 2009;52:1699–1708.[CrossRef][Medline]
- Jonasson JM, Ljung R, Talbäck M et al. Insulin glargine use and short-term incidence of malignancies—a population-based follow-up study in Sweden. Diabetologia 2009;52:1745–1754.[CrossRef][Medline]
- SDRN Epidemiology Group. Use of Insulin Glargine and Cancer Incidence in Scotland: A Study From the Scottish Diabetes Research Network Epidemiology Group. 2009. Available at http://webcast.easd.org/press/glargine/download/090818Colhounacceptedpaperadjusted.pdf. accessed July 10, 2009.
- Currie CJ, Poole CD, Gale EA. The influence of glucose-lowering therapies on cancer risk in type 2 diabetes. Diabetologia 2009;52:1766–1777.[CrossRef][Medline]
- Calle EE, Rodriguez C, Walker-Thurmond K et al. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 2003;348:1625–1638.
[Abstract/Free Full Text] - Renehan AG, Tyson M, Egger M et al. Body-mass index and incidence of cancer: A systematic review and meta-analysis of prospective observational studies. Lancet 2008;371:569–578.[Medline]
- Erbel S, Reers C, Eckstein VW et al. Proliferation of colo-357 pancreatic carcinoma cells and survival of patients with pancreatic carcinoma are not altered by insulin glargine. Diabetes Care 2008;31:1105–1111.
[Abstract/Free Full Text] - Rosenstock J, Fonseca V, McGill JB et al. Similar risk of malignancy with insulin glargine and neutral protamine Hagedorn (NPH) insulin in patients with type 2 diabetes: Findings from a 5 year randomised, open-label study. Diabetologia 2009;52:1971–1973.[CrossRef][Medline]
- McMaster University. Research Team Responds to Concerns About International Insulin Drug Trial. McMaster University news release. 8 10, 2009. Available at http://fhs.mcmaster.ca/main/news/news_2009/origin_trial.html. accessed August 22, 2009.
- Grill M. Dünger für Krebszellen. Der Spiegel 2009;104:104–107.
- Rummel A. Erhöht Lantus das Krebsrisiko? TV—Das Erste 7 13 2009;11–16.
- German Society for Haematology and Oncology. Keine Steigerung des Krebsrisikos durch Insulin Lantus [press release]. 7 15 2009.
- Institute for Quality and Efficiency in Health Care. IQWiG. Abschlussbericht: Langwirksame Insulinanaloga zur Behandlung des Diabetes mellitus Typ 2. Germany: Cologne, 2009. Available at http://www.iqwig.de/download/A05–03_Abschlussbericht_Langwirksame_Insulinanaloga_bei_Diabetes_mellitus_Typ_2_V1.1.pdf. accessed August 22, 2009.
- Horvath K, Jeitler K, Berghold A et al. Long-acting insulin analogues versus NPH insulin (human isophane insulin) for type 2 diabetes mellitus. Cochrane Database Syst Rev 2007;(2):CD005613.
- Wood L, Egger M, Gluud LL et al. Empirical evidence of bias in treatment effect estimates in controlled trials with different interventions and outcomes: Meta-epidemiological study. BMJ 2008;336:601–605.
[Abstract/Free Full Text] - European Medicines Agency Press Office. Press Release: European Medicines Agency Update on Safety of Insulin Glargine. London. 7 2009. 23. Available at http://www.emea.europa.eu/humandocs/PDFs/EPAR/Lantus/47063209en.pdf. accessed August 4, 2009.
This article has been cited by other articles:
 |
|
 |
|
  J. M. de Miguel-Yanes and J. B. Meigs When "Flawed" Translates into "Flood": The Unproven Association Between Cancer Incidence and Glargine Insulin Therapy Oncologist, December 1, 2009; 14(12): 1175 - 1177. [Full Text] [PDF]
|
|
|
|
|
|
C. R. Chong and B. A. Chabner Mysterious Metformin Oncologist, December 1, 2009; 14(12): 1178 - 1181. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| THE ONCOLOGIST | STEM CELLS | CME | ALPHAMED PRESS JOURNALS |