For people with type 2 diabetes, quitting smoking may lead to a worsening of glycemic control unrelated to weight gain, a new study suggests.

The findings, from a large UK primary-care database, were published online April 29 in Lancet Diabetes & Endocrinology by Deborah Lycett, PhD, RD, clinical dietitian and principal lecturer in nutrition and dietetics at Coventry University, United Kingdom, and colleagues.
Of 10,692 type 2 diabetes patients who were current smokers on January 1, 2005, the 29% (3131) who subsequently quit for at least a year had an average 0.21% increase in their HbA1c levels (7.9% vs 7.7%) that wasn’t associated with changes in body weight or prescribing practices and that lasted 3 years post–smoking cessation.

“We suspected we would find the rise in HbA1c would have been explained by the weight gain that accompanies smoking cessation, but when we investigated this we didn’t find the evidence to support the theory in our data — which was surprising,” Dr Lycett told Medscape Medical News.

Findings Should Not Deter People From Quitting Smoking

However, the authors of an accompanying editorial question whether the study findings imply causation.

“We cannot infer from these results that stopping smoking causes increases in HbA1c concentrations because, despite adjustment for a range of clinical and demographic factors, the observational data presented might still be biased by residual confounding, both by indications for treatment and by lifestyle factors other than smoking status,” write Amy E Taylor, PhD, and colleagues, of the MRC Integrative Epidemiology Unit at the University of Bristol, United Kingdom.

Furthermore, severity of diabetes might affect patients’ success in stopping smoking, “so reverse causality cannot be ruled out,” they add.

Moreover, both the authors and the editorialists strongly emphasize that the findings should not deter clinicians from encouraging patients to quit smoking.
According to Dr Lycett, “These findings are about supporting patients to make a successful quit attempt, not to deter people from quitting. Rather, being aware of a rise in HbA1c allows both patients and clinicians to prepare for this and respond to it in order to have optimal diabetes control.” Such measures include optimizing statin and blood-pressure treatment and adjusting diabetes medications, she said.

Indeed, Dr Taylor and colleagues observe, “Establishing causality is unlikely to alter clinical messages about smoking cessation, because the benefits of quitting clearly outweigh any potential negative effects on health.”

Independent Effect

The figures analyzed by Dr Lycett and colleagues came from the Health Improvement Network (THIN) database, which includes electronic medical records from over 3.5 million patients in 546 UK primary-care practices.

The data set is fairly complete, since general practitioners are paid in part for their performance in caring for diabetes patients, including assessing their smoking status and encouraging quitting, as well as recording HbA1c and targeting a level less than 7.5%.

Adjustment for age, sex, diabetes duration, baseline body weight, statin prescription, and other factors did not significantly change the 0.21% difference in HbA_1c between those who quit smoking and those who didn’t, which disappeared after 3 years.

The patients who stopped smoking gained an average of 4.8 kg. But for every 1-kg increase in weight there was only a 0.008% rise in HbA_1c, “which would have a clinically negligible effect on HbA_1c for most quitters,” the authors say.

The investigators looked to see whether the rise in HbA_1c might be due to more medication being prescribed to smokers, but instead found the opposite: more of those who stopped smoking than continual smokers were on additional glucose-lowering medications at the end of follow-up than at the beginning.

For example, at the start of the study, 34.6% of continual smokers vs 32.0% of quitters were on metformin monotherapy. By the end of follow-up, 29.5% of continual smokers vs 22.1% of quitters were on metformin alone. At study start, 13.2% of smokers and 15.9% of quitters had begun taking injectable glucose-lowering treatments, while by the end those proportions were 21.1% and 29.5%, respectively.

Dr Lycett told Medscape Medical News, “The deterioration in HbA_1c we found was small and temporary, over a period of 3 years, but the follow-up of the study was not long enough to determine what the long-term impact of this would be on macrovascular and microvascular complications.”

Teasing Out Causality

The lack of association with weight gain leaves the explanation for the findings unclear.

Dr Lycett said, “Even if weight gain is not the explanation behind this, it is possible that dietary change is, as we know that the preference for sweet-tasting food increases when people stop smoking, so this could potentially raise the glycemic load of their diet and their HbA_1c. However, we as yet have no data to support this.”

In their editorial, Dr Taylor and colleagues suggest the use of causal inference methods to “provide more robust evidence” about the effects of smoking and smoking cessation on diabetes. Such methods might include assessment within the THIN database of physicians’ prescribing preferences for varenicline (Champix [UK], Pfizer) or nicotine-replacement therapy, since quit rates would be expected to be higher with varenicline.

Alternatively, they say, Mendelian randomization analysis using genetic variants related to smoking behavior could also help illuminate the impact of smoking on glycemic control, as well as reduce bias associated with body-weight–measurement error.

In any case, Dr Lycett told Medscape Medical News that the data on the cardiovascular benefits of quitting smoking for people with type 2 diabetes are well-established.

“This would suggest that the long-term benefits of quitting smoking definitely outweigh the temporary problems associated with it. However, it stands to reason that the optimal outcome would be to both quit smoking and to have good blood glucose control.”