Pseudonormal Diastolic Filling Unmasked With Glyceryl Trinitrate in Patients With Type 2 Diabetes With Poor Metabolic Control

Poirier et al. (1) report very interesting data regarding pseudonormal diastolic filling in men who are free of clinical heart disease and have well-controlled diabetes. Specifically, 28% of their subjects had a pseudonormal filling pattern, which was detected using a Valsalva maneuver and/or pulmonary venous Doppler (PVD). We conducted a similar study in 34 high-risk type 2 diabetic patients (16 men, 18 women) who were middle-aged (mean age 55.8 years), obese (BMI 30.6 kg/m²), and had poor metabolic control (HbA_1c 10.6%, fasting blood glucose 13.4 mmol/l). None of the patients had a history of heart failure or myocardial infarction, and we included patients with controlled hypertension. The patients underwent echocardiography methods similar to those used by Poirier et al. (1), although preload reduction was achieved with sublingual glyceryl trinitrate (GTN) (400 μg). PVD was suboptimal in one-third of the subjects and thus not included in the analysis.

Diastolic filling pattern was assessed at baseline, and 14 (41%) of the subjects had a normal filling pattern (E-to-A ratio 1.0–1.7, deceleration time 0.14–0.23 s); the remainder had an abnormal filling pattern. However, after administration of GTN, all of the patients had an abnormal relaxation pattern; 20 (58.8%) were classified as having abnormal relaxation and 14 (41%) as having pseudonormal relaxation.

A total of six subjects in the abnormal relaxation group had left ventricular (LV) dilatation (M mode: LV end-diastolic pressure >58 mm). In all but one of these subjects, LV size was normal when adjusted for body surface area (BSA) LV end-diastolic dimension [LVEDD]/BSA <32 mm/m²). All patients had normal LV function as assessed by M mode fractional shortening (>25%) and subjective assessment of ejection fraction from the apical views. Compared with the abnormal relaxation group, the pseudonormal filling group was younger (52.6 vs. 58.1 years, P = 0.03) and had lower fasting blood glucose (12.1 vs. 14.3 mmol/l, P = 0.03). However, there were no differences in the levels of HbA_1c (10.3 vs. 10.8%, P = 0.14), duration of diabetes (6.8 vs. 7.9 years, P = 0.48), LV size (LVEDD: 50.4 vs. 54.1 mm, P = 0.16) or LV mass (143.9 vs. 156.6 g, P = 0.38).

Our group of patients had higher HbA_1c levels than those reported by Poirier et al. (1), none displayed a normal diastolic filling pattern, and 41% displayed a pseudonormal pattern in the absence of LV dilatation, hypertrophy, or systolic dysfunction. Our findings support the conclusions of Poirier et al. (1), highlighting the need for thorough echocardiographic evaluation in type 2 diabetes and, in particular, assessment of mitral filling under different loading conditions for thorough assessment of LV diastolic function.

In our study group, PVD was suboptimal in many subjects and thus may not be widely applicable in type 2 diabetic patients. Poirier et al. (1) found that while preload reduction always identified pseudonormal filling, PVD did not. Pseudonormal filling detected by preload reduction is related to LV diastolic pressure (2) and has been used to differentiate between true and pseudonormal filling patterns in both disease and healthy control subjects (3,4).

Traditionally, the focus of echocardiography has been to assess LV hypertrophy and systolic function. However, the presence of diastolic dysfunction in the setting of normal systolic function and no other structural heart disease may be important for long-term cardiovascular prognosis in diabetes. The pseudonormal group may represent a separate subgroup of patients who are at particular risk of developing a restrictive diabetic cardiomyopathy. A thorough evaluation using a longitudinal study is necessary to measure the natural time course of diastolic filling changes in diabetes, which would require the follow-up of a large cohort of patients.