September 7th, 2013

Visit-to-Visit Variability in Systolic BP Among Patients with Type 2 Diabetes


CardioExchange’s John Ryan interviews John Chalmers, Jun Hata, Hisatomi Arima, Peter M. Rothwell, and Mark Woodward about their new analysis using data from the ADVANCE trial.


The investigators analyzed data from 8811 patients with type 2 diabetes who did not die or experience major macrovascular or microvascular events within 2 years after randomization to perindopril + indapamide or matching placebo in the ADVANCE trial. During a median 2.4 years of follow-up after the 2-year visit, visit-to-visit systolic BP variability and maximum SBP each were independent risk factors for major macrovascular events (MI, stroke, CV death) and microvascular events (new or worsening nephropathy or retinopathy). The findings persisted after adjustment for mean SBP and other potentially confounding factors, although the association between maximum SBP and microvascular events did not persist in one adjusted model. SBP variability and maximum SBP were significantly associated with all-cause mortality, whereas mean SBP was not.


Ryan: The finding that visit-to-visit variability adds prognostic information beyond mean SBP seems to confirm findings from several previous studies. What key knowledge does your paper add?

Chalmers, Hata, Arima, Rothwell, and Woodward: The very new knowledge is that visit-to-visit SBP variability predicts the risk for microvascular complications over and above predicting the risk for death and macrovascular events. The other important finding is that these associations are present and strong in patients with type 2 diabetes.

Ryan: You discuss many potential mechanisms by which the variability might confer increased risk, including medication nonadherence, which seems plausible. Did the variability in SBP correlate with variability in other parameters that might relate to adherence but not to arterial stiffness (your putative mechanism for how variability relates to risk)?

Chalmers, Hata, Arima, Rothwell, and Woodward: We do not have specific information on nonadherence to medication, but we do have data on discontinuation of the randomized study medication—either perindopril/indapamide or placebo. In our sensitivity analyses, the effects of visit-to-visit variability remained significant after excluding patients who discontinued the study medication.

Ryan: What should clinicians do differently as a result of this study? Should they calculate visit-to-visit variability? If so, how do you recommend they do that?

Chalmers, Hata, Arima, Rothwell, and Woodward: These are possibly the most important lessons for the practicing clinician:

1. BP fluctuates widely from visit to visit and across home BP measurements.

2. The occasional high SBP reading (as reflected in the “maximum SBP,” a significant predictor of death and CV events) should prompt immediate action to improve BP control, rather than the more-typical response: “Let’s wait until the next visit to see what your BP is before we increase your medication.” That habit leads to very strong dependence on the mean SBP and ignores the fact that SBP variability constitutes another important and independent predictor or risk factor.

We do not recommend that the average practicing doctor try to calculate one of the various indices used in our formal analyses, such as the standard deviation or the coefficient of variation. Simple common sense should suffice.


Will the prognostic value of visit-to-visit variability in systolic BP influence how you manage hypertension in your patients with type 2 diabetes?

Comments are closed.