Caveat anicula! Beware of quiet little old ladies

It is conservatively estimated that between a third and half of patients hospitalised for decompensated heart failure have heart failure with preserved systolic function (HFPSF).1 Despite the prevalence of this syndrome, there are no robust, generally agreed diagnostic criteria for diastolic heart failure. International definitions of HFPSF2-4 define patients as having symptoms (shortness of breath, fatigue, orthopnea) or clinical signs of fluid retention (pulmonary, abdominal, or peripheral), and normal to near-normal systolic function on echocardiography (left ventricular ejection fraction [LVEF] > 45%),2 with clinical improvement in response to conventional treatment for chronic heart failure (CHF), if the diagnosis is in doubt.

Patients with HFPSF have a distinct demographic profile, aetiological background and pathophysiology.1,5 Compared with patients who have a low ejection fraction, previous studies suggest that those with HFPSF are generally older, are more often women, and are more likely to have CHF of hypertensive aetiology.6,7

HFPSF also has a different natural history compared with heart failure with left ventricular systolic dysfunction (LVSD), despite producing a similar symptom burden and mortality.1,8 The multiple comorbid conditions frequently associated with HFPSF contribute substantially to the risk of hospitalisation as HFPSF progresses.8,9

To date, most of the data on the epidemiology and natural history of HFPSF, and the associated disease burden, have come from clinical trials.7 Treatment for HFPSF remains largely empirical, as most drugs and devices that have been shown to reduce morbidity and mortality in heart failure have predominantly been tested in patients with LVSD.7

Accordingly, we undertook to characterise the demographics, pharmacotherapy, readmission rates and survival of patients with HFPSF from clinical data collected over a period of 10 years.

Admissions data and patient demographic information were acquired from hospital electronic records, with death data obtained from the National Death Index (developed and maintained by the Australian Institute of Health and Welfare). Comorbid conditions were derived from International Classification of Diseases, 10th revision (ICD-10) coding, and the medications prescribed from pharmacy dispensing data. Results of biochemical analyses were extracted using linked records from the hospital pathology service.

Box 1 shows that the mean age of patients in the HFPSF cohort was 2 years older than that of patients in the LVSD cohort.

Patients with HFPSF were much more likely to be women, and were more likely to have fewer social supports, less likely to be married and more likely to be widowed, less likely to live in their own home and more likely to be living in a nursing home (Box 1).

We undertook this study to characterise the demographic characteristics, pharmacotherapy, readmission rates and survival of patients with HFPSF from clinical data collected for patients admitted over a period of 10 years. The demographic characteristics of our HFPSF group closely mirrored those reported in other epidemiological studies. In particular, patients with HFPSF were more likely to be women and more likely to be older (age, > 75 years).6,7

Importantly, our study is the first to identify significant differences in the social environment of these patients. Specifically, patients with HFPSF were less likely to live in their own home, or have a spouse. Lack of social and carer support may have played a significant predisposing role in hospitalisation and readmissions.14,15

Patients with HFPSF had a higher prevalence of associated comorbid conditions, such as anaemia and atrial fibrillation, and such conditions have the potential to influence readmission and survival. Comorbid conditions may also have played a role in the therapeutic decision making of clinicians, potentially adversely impacting the intensity of therapy.16

Other studies have shown that patients with HFPSF were more likely to have hypertensive aetiology,7 but we found no significant difference in the prevalence of hypertension between the LVSD and the HFPSF groups. The higher echocardiographic prevalence of left ventricular hypertrophy among HFPSF patients suggests that this result may be due to variability in the coding of hypertension, rather than accurately reflecting the true historical prevalence of this condition in each cohort.

We observed that patients with HFPSF received less intensive pharmacological management than patients with LVSD. The evidence for pharmacotherapy in HFPSF is less rigorous than in LVSD but there are recognised studies supporting the importance of HFPSF pharmacotherapies.2-4 Patients with HFPSF were less likely to be prescribed an angiotensin-converting enzyme inhibitor, despite such agents having been shown to reduce hospitalisation rates and improve functional capacity in patients with HFPSF in randomised clinical trials.17

It is well appreciated that β-blockers have numerous theoretical benefits in patients with HFPSF, including lowering of heart rate (potentially improving diastolic filling and coronary blood flow, particularly during activity), reduction in myocardial oxygen demand and, by lowering the blood pressure, regression of left ventricular hypertrophy.18 Heart rate reduction is known to be particularly important in the treatment of pulmonary congestion as a result of diastolic heart failure secondary to ischaemia, and in patients with atrial fibrillation,19 and has recently been identified as an important determinant of β-blocker efficacy in the prevention of early mortality in LVSD.

As with β-blockers, atrial fibrillation may also have contributed to the similar prescription rates for digoxin in both patients with HFPSF and LVSD, despite concerns that have been raised about the use of digoxin in HFPSF. Classically, digitalis glycosides are known to increase intracellular calcium concentrations, potentially impairing myocardial relaxation and further worsening diastolic dysfunction. It is possible, however, that the sympatho-inhibitory, pro-parasympathetic and renin, angiotensin and aldosterone suppressing actions of digoxin are beneficial in HFPSF, but this remains to be proven in the clinical setting.20

We observed a higher rate of hospital readmission in the HFPSF cohort compared with the LVSD cohort. Previous reports, however, have identified readmission rates that were lower12,21,22 or comparable6,23,24 to those of patients with LVSD. Notably, our HFPSF cohort was older, with less social support, and a greater burden of comorbid conditions, so readmission cannot be attributed solely to differences in cardiac systolic function.

It has been suggested that HFPSF confers a better prognosis than heart failure with impaired systolic function in terms of morbidity and mortality. Studies that examined cohorts comparable to those in our study (mean age, > 65 years) have reported similar mortality rates among patients with HFPSF and LVSD.6,23,24 On the other hand, previous studies in younger populations (mean age, < 65 years) have shown better survival for HFPSF than LVSD.6,23,24 Whether such differences relate to differences in the relative malignancy of the underlying aetiology among younger LVSD patients, or are the result of an excess burden of morbidity and mortality related to comorbid illnesses and social isolation in the older HFPSF population remains uncertain.

In older patients, the clinical syndrome of CHF carries a uniformly poor prognosis regardless of the level of systolic function. Our findings, and those of similar studies, serve to heighten awareness of the prognostic impact of HFPSF, particularly among older patients.7

Numerous guidelines now indicate that Doppler-based parameters (eg, the ratio of peak early-diastolic transmitral flow velocity to peak early-diastolic mitral annular velocity [E/E' ratio] and pulmonary vein velocities) may be used to further clarify the diagnosis of diastolic impairment in the diagnosis of HFPSF. However, our study has shown a significant independent survival disadvantage, regardless of these Doppler-based parameters, when clinical findings of CHF are associated with normal systolic left ventricular function. It remains to be determined whether these Doppler echocardiographic parameters, coupled with novel diagnostic biomarkers such as brain natriuretic peptide, provide incremental prognostic value in the management of patients with HFPSF in routine clinical practice.25-28

Our study has a number of limitations that require comment. These data are based on a cohort of patients from a single tertiary institution. Despite this, the clinical data available for this cohort were substantially more detailed than reported previously. The index admission was that recorded at the tertiary hospital of interest; however, readmissions data from all tertiary and some regional hospitals throughout the state were used to ensure the maximum possible readmissions data were captured.7

The clinical underutilisation of echocardiography was a considerable limitation of this study. Capturing echocardiography data was complicated by patients having been discharged before an echocardiogram was performed, and those who had used private imaging services after discharge making records inaccessible. Although there may have been a bias towards patients with more severe symptoms having echocardiogram data available, this is unlikely to alter the findings of our study with regard to the characteristics of those with impaired compared with preserved systolic function.

Prescribing data reflected what patients were dispensed from the hospital pharmacy, but do not capture medications on ward stock, such as loop diuretics, and do not take into account potential changes to pharmacotherapy after discharge.

In the context of less compelling evidence for aggressive cardiopharmacological management, our study shows that these predominantly older female patients with HFPSF had less social support and higher readmission rates for associated comorbid illnesses. We propose that the burden of reduced survival in HFPSF may relate more to comorbid conditions than suboptimal cardiac management. Therefore, we would warn clinicians to caveat anicula! — beware of the little old lady with a “normal” echocardiogram.

3 Comparison of survival among patients with left ventricular systolic dysfunction (LVSD) and heart failure with preserved systolic function (HFPSF)

Year*	0	1	2	3	4	5	6	7	8	9
Surviving patients
LVSD	564	415	330	270	216	174	135	88	31	3
HFPSF	189	147	114	85	62	48	36	23	12	2
* Time from the first matching echocardiogram until the last date of follow-up.