The complications of diabetes are generally divided into two groups, macrovascular and microvascular. The former include heart disease as manifest mostly by heart attacks and strokes. The latter involve problems with vision, kidneys, and peripheral circulation. Advanced microvascular disease results in blue feet, gangrene, amputations, blindness and kidney failure.
The conventional wisdom is that by controlling blood sugar, type 2 diabetics can significantly decrease the risk of microvascular complications. In fact because there is growing awareness that even intensive glucose control does not reduce macrovascular risk,1 the emphasis is shifting to microvascular. That is all there is left. Expert opinion generally considers the evidence convincing. In the past five years, several studies have reported results that seem at variance with the view of microvascular benefits accruing from intensive blood sugar control. A damage control paper appeared reviewing the major negative studies that attempted to neutralize the impact of these results on both clinicians who have little else to offer and the pharmaceutical industry which has a strong interest in maintaining the status quo.2 Many of the authors had strong ties with the pharmaceutical industry. It is thus of interest to examine of what the microvascular trial evidence actually consists and to what extent can it be regarded as clinically significant.
One of the most frequently cited trials is the United Kingdom Prospective Diabetes Study (UKPDS).3,4 This study actually started in 1983 and reported 10 year results finally in 1998. Over this period, the primary and other endpoints were changed a number of times and the announced termination date repeatedly extended. Furthermore, the study was not blinded for the investigators and the results of interim analysis were available. This study protocol is obviously open to bias. To quote one critic,"It seems that the authors continued the study until they obtained a result that was significant without adjusting for repeatedly looking at the data".5 An investigator involved in the trial commented in 2008 that "UKPDS broke almost all the rules of trial design. We are taught to believe that a study protocol should be predetermined and set in stone, but this study went to the other extreme, elevating the ad hoc into an art form".6 Thus the key study appears flawed. At the end of the study period, participants were followed for a number of years but as regards glucose control, apparently did whatever they pleased. Thus, one can argue that the only results meriting attention are those for the actual study, and as pointed out above, the validity of the actual trial has been questioned.
UKPDS actually included several concurrent studies. UKPDS-34 compared intensive treatment with what was termed conventional treatment, evidentially starting with diet followed by drugs.3 The conventional treatment involved maintaining blood sugar below 15 mmol/L (270 mg/dL). When marked hyperglycemia occurred drug treatment was initiated which included a sulphonylurea or insulin. The intensive treatment group had as its target fasting blood glucose of < 6 mmol/L. Metformin or one or more sulphonylurea were used as well as insulin. Changes in the drug protocols occurred during the trial, leading to confusion which makes it difficult to interpret the published results. Upon comparison of the intensive vs. conventional groups at presumably the end of the study, strongly insignificant microvascular results were obtained for death from peripheral vascular disease, amputation, death from kidney disease, kidney failure, blindness in one eye, need for photocoagulation, cataract extraction and ocular haemorrhage. The same picture emerged when sulphonylurea plus metformin were compared with sulphonylurea alone. Except for an initial drop in fasting glucose and HbA1c, both the intensive and conventional groups experience steady progression of these two markers over the course of the trial and they were always diagnostic of diabetes. UKPDS-34 had 21 individual endpoints, both microvascular and microvascular. With that number of endpoints, the random chance of obtaining a false positive is 66%.7
In UKPDS-33,4 which reported at the same time as UKPDS-34, intensive drug therapy with sulphonylureas or insulin was compared with convention treatment defined above, and seven of eight microvascular endpoints yielded results showing no significant impact. Only the need for retinal photocoagulation resulted in a significant microvascular relative risk reduction and this was close to being insignificant and did not agree with the results in UKPDS-34, which were negative. Photocoagulation is also considered by some to be a soft endpoint since the judgement of the attending physician is involved.
The investigators also provided graphs of fasting blood glucose and HbA1c vs. time for conventional and metformin use and for conventional and in intensive therapy. Drug therapy caused a sharp drop in both at about year 1 and then values increased in parallel with almost the same rate after 6 years. In other words, the treatments used did not have much influence on the relentless progression of the disease except allowing it to progress with somewhat lower fasting glucose and HbA1c levels. This is part of what mainstream medicine calls strong evidence of benefit in the context of preventing microvascular complications.
The large ACCORD study reporting in 2008 found no macrovascular benefits associated with intensive treatment of hyperglycemia in type 2 diabetics compared with standard treatment.8 ACCORD was stopped after a mean treatment time of about 3.5 years of follow-up because of increased mortality in the intensive compared to the standard therapy groups. The target for intensive therapy was HbA1c <6% whereas the standards therapy aimed at 7.0-7.9%. All subjects started with HbA1c levels of = 7.5%.At baseline, both groups were on anti-hyperglycemic drugs. Only macrovascular events were reported in the 2008 paper and no benefits were found for intensive vs. non-intensive treatment. After the decision was made to stop the trial, the intensive treatment group was switched to standard therapy. In a report covering the posts termination follow-up out to 5 years, it was concluded that the strategy of intensive therapy could not be recommended for the high-risk, advanced type 2 diabetics involved in the trial.
There was a microvascular ACCORD eye study which started at the same time as the main trial and involved 4 years of follow-up.9 The report in 2010 was corrected in 2011. The revised results for the impact of intensive vs. standard glucose control on diabetic disorders of the retina were not statistically significant, nor was the original published result.
A more detailed account of the parallel microvascular studies in the ACCORD trial was published in 2010.10 The investigators report results up to the trial termination and for an additional post-trial follow-up. Since having a study interrupted and the treated and controls subsequently given the same therapy with continued follow-up confuses the issue, only the actual trial results will be discussed. No eye related endpoints were found to have significant benefit in the comparison of the standard and intensive groups. Only the incidence of micro- and macro albuminuria (kidney function marker) had significant relative risk reductions for the intensive vs. the standard therapy. The NNT provided in the paper were 39 and 73 respectively (97.4 and 98.6% failed to benefit). However, the clinical significance of these results is somewhat diluted by the observation that both before and after the trial termination, there was a steady increase in blood creatinine levels, another kidney function marker. Since there were no differences in the behavior of this marker between the intensive and standard groups, it appears that the intensive therapy had no beneficial influence on the actual progression of kidney disease in these patients. The authors actually point out that given the 15 endpoints studied, there was a 54% chance of randomly obtaining a false positive result.
Next we come to the ADVANCE study, which was one of two reporting in 2008 that shook up the world of diabetology.11 ACCORD was the other. ADVANCE was a randomized trial lasting 5 years which aimed in the treated group to achieve a target HbA1c of 6.5% with intensive glucose control using Gliclazide. Controls followed conventional guidelines for treatment. The control group maintained an HbA1c of close to baseline (7.4%) whereas the intensive group dropped to 7.0% within 6 months and 6.5% within 36 months which was then maintained. Contrary to the UKPDS, no effect on visual complications was found and the only benefit for microvascular complications involved kidney disease. Among the primary endpoints, the benefit regarding new or worsening kidney disease had a number needed to treat (NNT) of 90 indicating that 98.9% of the subjects derived no benefit. For the secondary endpoints, the benefit associated with new-onset microalbuminuria had an NNT of 50 (98% no benefit). There were 14 secondary endpoints designated, and thus there was a 51% chance of randomly obtaining a false positive. The results were non-significant for peripheral vascular events, and new or worsening neuropathy.
A study which also looked at microvascular benefits examined the efficacy of metformin vs. a control group of patients who had withdrawn from the drug and were treated by a non-drug approach.12 The urinary albumin/creatinine ratio (kidney function) and the progression of eye problems associated with the retina were included in the outcomes and no benefit was found in the comparison of treated vs. untreated patients for any endpoint. The trial was randomized and lasted 4 years, and in terms of the control group, was similar to a placebo controlled trial.
A smaller randomized trial, the Veterans Affairs Diabetes Trial (VADT) which reported in 2009 involved a comparison between intensive and standard glucose control in type 2 diabetics with suboptimal response to therapy. In this study 1791 men, mean age 60, were randomized to two groups with a median follow-up of 5.6 years. Median HbA1c dropped from 8.4% in the standard therapy group to 6.9% in the intensive group. The only microvascular complication to show an indication of benefit from intensive glucose control was the progression to kidney disease. But statistically significant increased incidence of nephropathy was only seen in an analysis when micro- and macroalbuminuria were combined.13
Prompted by the ADVANCE and VADT results and earlier studies related to kidney complications, a study involving meta-analysis was reported in 2012. It investigated intensive vs. standard glucose control to determine if this decreased significant renal clinical outcomes such as doubling of serum creatinine levels, end-stage renal disease or death from kidney disease during the years of follow-up. No evidence of benefit was found when 7 trials with follow-up from 2 to 15 years were included.14
Rightly or wrongly, mainstream medicine regards the meta-analysis as the path to truth and enlightenment. What do other meta-analyses say about the treatment of the microvascular complications of type 2 diabetes? Hemmingsen et al included both composite microvascular outcomes and disorders of the retina and kidney.15 This meta-analysis found no benefit for lowering the risk of kidney disease. For composite microvascular outcomes, 98.5% failed to benefit, whereas for disorders of the retina, the results was 97.7%. Boussageon et al published a meta-analysis in 2012 which included only studies where metformin could be compared with either a placebo, diet or other drug interventions. The analysis examined peripheral vascular complications, amputation, and a microvascular composite and found no significant benefit.16
Thus, it appears that the evidence needed to back up the view that blood glucose control for reducing the risk of microvascular complications in diabetics is almost entirely absent. Some studies find suggestions of a modest benefit with large NNT. Furthermore, one does not know how much weight to give the photocoagulation result in UPDPS because of the flaws described above. The current position of mainstream medicine, rather than resting on a convincing base of evidence, appears to be influenced by there being nothing else to offer patients with type 2 diabetes. The patients see improvements in their glucose control are naturally convinced that there is benefit. They believe they are preventing complications. This is why, form their point of view, they are taking the drugs. Unfortunately, the treatment does not appear to provide significant benefit.
The reader is referred to the November IHN where a diet-based approach to type 2 diabetes is discussed which accomplishes near or total reversal that appears to be durable (drugs or insulin no longer needed). Most patients on the diet were successful. The data presented above indicate that most treated with drugs do not benefit. They are on drugs for life and mostly just get worse. No drug treatment can accomplish what the diabetes reversal diet protocol does and the benefit occurs in 8 weeks. The contrast to the relentless progression of this disorder in patients treated with anti-hyperglycemic drugs, even to low, near normal targets of fasting glucose and HbA1c, is striking.
We are not talking about some rare disease but a worldwide epidemic that, along with the horror of the complications, threatens to financially bring down so-called health care systems and is encompassing younger individuals each year. Recall that type 2 diabetes is no longer called "adult onset".
This review completes the discussions in IHN on the absence of efficacy associated with drug treatment of type 2 diabetes in the context of diabetic complications.