analysis and trial sequential analysis of randomised clinical trials
Risk ratios with 95% confidence intervals were estimated with fixed and random effects models.
Results Fourteen clinical trials that randomised 28614 participants with type 2 diabetes (15269 to intensive control and 13345 to conventional control) were included. Intensive glycaemic control did not significantly affect the relative risks of all cause (1.02, 95% confidence interval 0.91 to 1.13; 28359 participants, 12 trials) or cardiovascular mortality (1.11, 0.92 to 1.35; 28359 participants, 12 trials). Trial sequential analyses rejected a relative risk reduction above 10% for all cause mortality and showed insufficient data on cardiovascular mortality. The risk of non fatal myocardial infarction may be reduced (relative risk 0.85, 0.76 to 0.95; P=0.004; 28111 participants, 8 trials), but this finding was not confirmed in trial sequential analysis. Intensive glycaemic control showed a reduction of the relative risks for the composite microvascular outcome (0.88, 0.79 to 0.97; P=0.01; 25600 participants, 3 trials) and retinopathy (0.80, 0.67 to 0.94; P=0.009; 10793 participants, 7 trials), but trial sequential analyses showed that sufficient evidence had not yet been reached. The estimate of an effect on the risk of nephropathy (relative risk 0.83,
van cleef and arpels perlee fake bracelet, 0.64 to 1.06; 27769 participants, 8 trials) was not statistically significant. The risk of severe hypoglycaemia was significantly increased when intensive glycaemic control was targeted (relative risk 2.39, 1.71 to 3.34; 27844 participants, 9 trials); trial sequential analysis supported a 30% increased relative risk of severe hypoglycaemia.
Conclusion Intensive glycaemic control does not seem to reduce all cause mortality in patients with type 2 diabetes. Data available from randomised clinical trials remain insufficient to prove or refute a relative risk reduction for cardiovascular mortality, non fatal myocardial infarction, composite microvascular complications, or retinopathy at a magnitude of 10%. Intensive glycaemic control increases the relative risk of severe hypoglycaemia by 30%.
IntroductionObservational studies suggest an association between the extent of hyperglycaemia and the risk of death and of macrovascular and microvascular disease in patients with type 2 diabetes.1 2 3 Three recent randomised clinical trials in patients with type 2 diabetes were not able to detect (or reject the possibility of) reduced cardiovascular disease or mortality with intensive compared with conventional glycaemic control.4 5 6 Worries arose as the results from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial in 2008 showed increased all cause mortality and cardiovascular mortality in the intensive treatment group compared with conventional treatment.4 The increased mortality led to early termination of the ACCORD trial.4 On the other hand, randomised clinical trials have indicated a beneficial effect on microvascular complications of intensive versus conventional glycaemic control in patients with type 2 diabetes. However, inconsistencies exist among the trials with respect to which type of microvascular complications are prevented and the magnitude of the effect of intensive glycaemic control.5 7 8 The price of intensive glycaemic control may be an increased risk of hypoglycaemia. Achieving intensive glycaemic control in patients with type 2 diabetes requires enormous effort from the patient as well as resources from the healthcare system,
van cleef clover fake bracelet, particularly compared with the well documented beneficial effects of lipid and blood pressure lowering treatment.9
The definition of intensive glycaemic control varies among trials and guidelines. The ACCORD trial and the Veterans Affairs Diabetes Trial (VADT) used a target of glycated haemoglobin A1c (HbA1c) below 6.0% for intensive glycaemic control compared with a target of HbA1c below 6.5% in the Action in Diabetes and Vascular Disease Preterax and Diamicron Modified Release Controlled Evaluation (ADVANCE) trial. The results from these trials have created a debate about the optimal choice of glycaemic target. The American Diabetes Association recommends an HbA1c level of less than 7.0% as the standard glycaemic treatment goal, whereas the International Diabetes Federation recommends a level of less than 6.5%.10 11 12
In our published protocol, we predefined inclusion of all trials comparing patients treated to a specific target for intensive glycaemic control with patients treated to a conventional but higher glycaemic target.13 The difference in treatment strategy between the groups was clearly defined either as values of HbA1c or as intensifying glycaemic control. The intensive glycaemic targets varied across the trials, but the trials compared the results of trying to achieve a distinct lower target with those of aiming for a higher one. We believe that the existence of a "gold threshold" target remains to be established and that the hypothesis so far has been that targeting/lowering the HbA1c may have a beneficial effect along the entire scale of measurements of HbA1c unless hypoglycaemia is reached. In this sense, we have included all trials comparing an intensified glycaemic target with a more "relaxed" glycaemic target, often reflecting usual clinical practice for a given place and time.
This systematic review reanalyses current evidence of the effect of targeting intensive glycaemic control on all cause mortality, cardiovascular mortality, cardiovascular disease, and microvascular disease in patients with type 2 diabetes. We consider the effects of intensive glycaemic control irrespective of differences among trials in individual targets or achieved glycaemic control.13MethodsThis review follows the recommendations of the Cochrane Collaboration.15 It is based on our published Cochrane protocol.13
We included all randomised trials that compared the targeting of intensive glycaemic versus conventional glycaemic control in patients with type 2 diabetes.13 14 We analysed trials according to the setting of the intensive glycaemic intervention. We analysed trials of targeting intensive glycaemic control in patients without acute events at entry or without concomitant treatments targeting other cardiovascular risk factors as "trials exclusively dealing with glycaemic control in usual care setting."13 The data in the review reported here are from this group of trials, representing 28614 (95%) of 29986 participants included in our review.13 14 We excluded three trials assessing multimodal interventions,16 17 18 as well as three trials assessing intensive glycaemic control as part of an acute intervention.19 20 21 For the vast majority of estimated effects of intervention, these exclusions did not cause noticeable changes.13
We analysed trials of targeting intensive glycaemic control as part of an acute intervention and trials with multimodal interventions separately.13 14 We also did an overall meta analysis combining data from all included trials irrespective of the setting in which intensive glycaemic control was applied.13 14 We refer only to data from the analyses of trials exclusively dealing with glycaemic control in usual care setting in this paper, but the Cochrane version gives a full presentation.14
Search strategyWe did a search in the Cochrane Library, Medline, Embase, Science Citation Index Expanded,
alhambra bracelet knock off van cleef, LILACS, and CINAHL in December 2010 for randomised clinical trials of targeting intensive glycaemic control versus targeting conventional glycaemic control in patients with type 2 diabetes. Web appendix 1 describes the search terms and strategies for each database. We also searched abstracts presented at the American Diabetes Association and European Association for the Study of Diabetes congresses. We contacted relevant drug companies and the US Food and Drug Administration for unpublished randomised trials relevant to our review. We searched reference lists of included trials and (systematic) reviews, meta analyses, and health technology assessment reports. We did internet searches for all trials to find additional information about the included trials. We included a trial if it was a randomised clinical trial, compared targeting intensive glycaemic control versus targeting conventional glycaemic control, and was done in patients with type 2 diabetes. We included trials irrespective of duration, language, publication status, or predefined outcomes. Web appendix 2 gives details. Translators extracted data from all relevant non English articles.
We extracted data on several baseline characteristics of the participants (such as age, duration of disease, HbA1c) and outcomes.
When we identified more than one publication of an original trial, we assessed these together to maximise data collection.
Statistical analysisWe used Review Manager version 5.0.25 for statistical analysis.22 We summarised data on all cause mortality, cardiovascular mortality, non fatal myocardial infarction, and severe hypoglycaemia statistically as relative risks with 95% confidence intervals. We used both a random effects model and a fixed effect model.23 24 In case of discrepancy between the two models, we report both results; otherwise, we report the random effects model.
We examined heterogeneity with the I2 statistic, quantifying the proportion of between trial variance to the sum of the between trial variance and a common sampling error.25 We graded values of I2 between 0% to 40% as "heterogeneity might not be important," values between 30% and 60% as "moderate heterogeneity," values between 50% and 90% as "substantial heterogeneity," and values between 75% and 100% as "considerable heterogeneity."25 When we found heterogeneity, we attempted to determine potential reasons by examining characteristics of individual trials.
We did subgroup analyses stratifying trials according to risk of bias, length of study, diagnostic criteria for type 2 diabetes, language of publication, and source of funding for all cause mortality, cardiovascular mortality, and non fatal myocardial infarction.
We did trial sequential analyses.26 27 This is similar to interim analyses in a single trial, where monitoring boundaries are used to decide whether a trial could be terminated early when a P value is sufficiently small to show the anticipated effect. Because no reason exists why the standards for a meta analysis should be less rigorous than those for a single trial, analogous trial sequential monitoring boundaries can be applied to meta analysis.28 29 30 Cumulative meta analyses of trials are at risk of producing random errors because of sparse data and repetitive testing of accumulating data when the required information size (analogous to the sample size of an optimally powered clinical trial) has not been met. Trial sequential analysis depends on the quantification of the required information size (the meta analysis sample size). In this context, the smaller the required information size the more lenient the trial sequential monitoring boundaries are and, accordingly, the more lenient the criteria for statistical significance will be. We calculated a heterogeneity (I) adjusted required information size. We did the trial sequential analyses with an intention to maintain an overall 5% risk of a type I error, which is the standard in most meta analyses and systematic reviews. On the basis of pre determined criteria,13 we initially calculated the required information size to detect or reject an intervention effect of a 10% relative risk reduction with a risk of a type II error of 20% (power of 80%). We chose a 10% relative risk reduction equivalent to a number needed to treat of approximately 100 patients, because even this decrease in mortality is likely to be clinically meaningful. For severe hypoglycaemia, however, we chose a 30% increase in relative risk equivalent to a number needed to harm of 50. We also provide the 95% confidence intervals adjusted for sparse data and repetitive testing, which we describe as the trial sequential analysis adjusted 95% confidence intervals.
ResultsResults of the search and trial,
van cleef and arpels clover knock off bracelet, participant, and intervention characteristicsFigure 1 summarises the results of the search. We excluded 42 references after further evaluation. The main reasons for exclusion were that the trial was not randomised (11 references),31 32 33 34 35 36 37 38 39 40 41 participants were not patients with type 2 diabetes or we could not separate data on patients with type 2 diabetes (four references),42 43 44 45 or no predefined differences in glycaemic intervention target existed (16 references).46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 In addition, we excluded trials that assessed intensive glycaemic control as a part of an acute intervention (five references, three trials) or had concomitant targeting of several cardiovascular risk factors in the glycaemic intervention arm (six references, three trials).16 17 18 19 20 21 62 63 64 65 66 Table 1 gives a list of excluded trials.
Fig 1 diagram of identification of randomised trials for inclusion
Two of the trials had a factorial design.4 5 The UK Prospective Diabetes Study (UKPDS) had a substudy in which some of the participants were randomised to intensive blood pressure control versus conventional blood pressure control.102 The University Group Diabetes Program (UGDP) randomised participants to five different treatment regimens.88 We chose to report the "insulin variable" group as the intensive group and the "insulin standard" group as the conventional group.
The Kumamoto trial had a planned length of intervention of six years.7 Only two of the included 110 participants changed their glycaemic intervention regimen after the predefined intervention period. The trial therefore continued on the initiative of the participants. We have reported all outcomes in this analysis after 10 years of follow up, except for severe hypoglycaemia (reported after eight years of follow up).95
The Wall