Home / Resources / Articles / Can HbA1c Variability Predict Diabetic Complications?

Can HbA1c Variability Predict Diabetic Complications?

Jul 6, 2021
Editor: Steve Freed, R.PH., CDE

Author: Macrina Ghali, PharmD Candidate 2021, Florida A&M University, College of Pharmacy and Pharmaceutical Sciences

A study was conducted to assess if HbA1c variability impacts the incidence of diabetic complications, all-cause and cardiovascular mortality.  

Increased variability in HbA1c has been associated with higher rates of diabetic complications and mortality. These significant fluctuations in HbA1c may increase mortality risk, especially in higher-risk populations such as geriatric patients. It is also unknown whether periods of hypoglycemia have negative impacts on organ function. Previous clinical trials have evaluated intensive glycemic control and found that more stringent goals reduce mortality and cardiovascular complications. However, these studies excluded patients with significant comorbidities. A retrospective observational study was conducted in Hong Kong, evaluating hypoglycemia, HbA1c variability, and mortality to determine the predictive power of HbA1c value and its association with diabetes prognosis.  


The cohort included 3,424 patients; however, only 3,127 patients met the criteria of having at least three HbA1c measurements. The primary outcomes were all-cause and cardiovascular mortality. Secondary outcomes included diabetic complications such as neurological, ophthalmological, and renal complications, microalbuminuria and macroalbuminuria, peripheral vascular disease, stroke, transient ischemic attack, atrial fibrillation, sudden cardiac death, and diabetic ketoacidosis (DKA) or hyperosmotic hyperglycemic state (HHS) or coma.  The average age was 63 years, and the most common comorbidities in the population were hypertension (24.6%), ischemic heart disease (15%), and stroke (11%). The average daily insulin was 20.8 units, and most patients had either metformin (1300 patients) or a sulfonylurea (1300 patients) as antihyperglycemic therapy.  

The variability in HbA1c was evaluated using mean, standard deviation (S.D.), root mean square (RMS), and coefficient of variation (CV). Researchers defined the HbA1c variability score (HVC), which was expressed as a percentage, as the number of HbA1c measurements >0.5% of the last reading divided by the total number of HbA1c measures. Statistical analysis was conducted using logistic regression to identify significant predictors of different outcomes, while Cox regression was utilized to determine the predictive value of HbA1c variability for the primary outcome. Researchers determined statistical significance if the P-value was less than 0.05. Furthermore, Kaplan-Meier curves analyzed the time-to-death for all-cause mortality.   

Both Cox and logistic regression found the baseline, mean, and RMS of baseline HbA1c and HbA1c variability to be negative predictors of all-cause mortality. At the same time, the SD, CV, and HVS were all positive predictors. P values for all Cox and logistic regression predictors were less than 0.001 and classified as statistically significant. For cardiovascular mortality and time-to-death, baseline HbA1c (P=0.005 for cardiovascular mortality and P=0.009 for time-to-death) was a negative predictor, while SD, CV, and HVS were positive predictors. P values for the positive predictors were less than 0.001.  

Researchers used cut-off values of 7.3% and 6.8% for HbA1c values and cut-off values of 0.86 and 0.88 for HbA1c variability in all-cause mortality and cardiovascular mortality risk predictions. The lower mean HbA1c group was associated with significantly shorter time-to-death for all-cause mortality (P<0.001) but not for cardiovascular mortality (P=0.920). In contrast, both time-till-death for all-cause mortality cardiovascular mortality was significantly shorter in the high HbA1c variability group (P<0.001 for both). Furthermore, hypoglycemia incidence was a favorable mortality and time-till-death predictor for all-cause mortality and a positive predictor for time-till-death in cardiovascular mortality. Similarly, both baseline HbA1c and HbA1c variability were positive predictors for the development of DKA, HHS, or coma, neurological or ophthalmological diabetic complications, microalbuminuria, and renal diabetic complications.  

This study showed that both high and low HbA1c values and variability could predict mortality and diabetic complications. Patients with a low HbA1c were associated with adverse outcomes, and frequent hypoglycemia episodes increased all-cause and cardiovascular mortality.  Although this study offers evidence for the utilization of HbA1c variability, the different parameters for calculations remain unclear and prevent its standardized use. This study’s limitations include its primarily Chinese population, preventing generalizability, and its observational nature, impacting data collection. In conclusion, this study highlights the negative impacts of intermittent hypoglycemia on mortality. It also demonstrates the possible relationship between HbA1c variability, prognosis, and risk of diabetic complications in patients with diabetes.   

Practice Pearls 

  • High HbA1c variability increased the risk of all-cause and cardiovascular mortality, in addition to the incidence of diabetic complications.  
  • Increased mortality was seen with more frequent hypoglycemia, which may prompt less stringent glycemic goals in specific populations.  
  • The pathophysiology of intermittent hypoglycemia on diabetic complications is not clearly understood and will need to be evaluated. 


Lee, S., Liu, T., Zhou, J. et al. Predictions of diabetes complications and mortality using hba1c variability: a 10-year observational cohort study. Acta Diabetol 58, 171–180 (2021).  

Gorst C, Kwok CS, Aslam S, et al. (2015) Long-term glycemic variability and risk of adverse outcomes: a systematic review and meta-analysis. Diabetes Care 38(12):2354–2369 


Macrina Ghali, PharmD Candidate 2021, Florida A&M University, College of Pharmacy and Pharmaceutical Sciences