The main results in this study are the following:
- There were few investigations of HF related to etiology and laboratory data taking place during the geriatric care episodes. Two patients were assessed with ECHO. One in five patients had their NT-pro-BNP followed. No NYHA classifications were documented.
- At admission, 83% of the HFrEF patients were treated with BB and ACEI/ARB, according to guidelines. Seventy percent of HFmrEF and 50% of HFpEF had similar treatment. The number of HFrEF patients treated with BB and ACEI/ARB decreased to 71, 63 and 49% respectively during the care episode.
- At discharge, 21% of those with HFrEF reached the target dose of ACEI/ARB. There had been no significant changes during the care episode concerning total prescription of ACEI/ARB or BB. There were significantly more patients treated with spironolactone (28 vs 19%) compared with admission and furosemide (83 vs 55%) compared with 14 days prior to admission. Only five patients were treated with target doses of both ACEI/ARB and BB.
- Those who had received more spironolactone (either new prescriptions or increased doses) had during the care episode greater weight loss, longer hospital stays, more assessments of NT-pro-BNP and were at discharge more often referred to nursing home.
- The majority (68%) of HF patients was discharged to own home with follow-up at primary care. Information on etiology of HF, when known, was rarely transferred in the referral from geriatricians to primary care physicians.
Investigation
The patients in the study were to a limited degree further investigated during the care episode for HF. There may be several reasons for these modest investigative actions. First, there may not be a need for investigations if such activities had been recently performed. However, only 30 patients (22%) had had both an ECHO and a NT-pro-BNP during the last 90 days prior to admission for treatment of worsening HF. Since the patients were high consumers of inpatient care and were admitted due to acutely deteriorated HF, further investigations should have been indicated.
Other explanations for the lack of performing ECHO during the care episode may be that it is not an articulated assignment for the clinic. Also, the cost of ECHO may decrease the ambitions. NT-pro-BNP, however, is cheap and accessible, but was still only performed for 27 of the 134 patients. A possible reason for the low use of NT-pro-BNP measurements may have been short hospital stays and thus less focus on further adjustment of treatment.
NYHA assessment is also cheap and in theory easy to perform. However, NYHA is harder to interpret among geriatric patients with a heavy burden of diseases, which may have an impact on the perceived usefulness and consequently actual assessment of NYHA. A limited walking ability or fatigue may have other explanations than HF, such as lung or orthopedic conditions or diseases.
The expressed mission from other caregivers when referring to geriatric care is normally to reduce HF symptoms and with that achieved good enough there may be no perceived need for further investigations. Moreover, the focus in recent years from the health care providers has been to reduce the length of in-patient hospital stay, which, in turn, may not leave enough time to perform investigations or optimize treatment.
Treatment changes during the care episode
International guidelines for treatment of HFrEF16 at NYHA stages II-IV recommend ACEI/ARB and BB as basic treatment and for certain indications with the addition of MRA. Diuretics are recommended for symptom relief. Treatment guidelines for HFpEF are less clear and recommend investigation and treatment of comorbidities and symptom relief. However, in clinical practice and clinical trials, only slightly fewer patients with HFpEF and HFmrEF, currently appear to receive diuretics, BB, MRAs and ACEIs or ARBs, compared with HFrEF patients26. This is not surprising, since comorbidities such as hypertension and atrial fibrillation, being indications for these drugs, are very common in all HF types.
In our study, 81% were treated at admission with BB, 66% with ACEI/ARB, 19% with spironolactone and 55% with diuretics (14 days prior to admission). Other studies have shown similar result with 79% BB27 and 69% ACEI/ARB28, but higher levels of use of spironolactone and diuretics (29% and 67% respectively)28.
During the studied care episodes, increases in spironolactone and furosemide prescriptions were made, resulting in a total increase of 28% of spironolactone and 83% of furosemide at discharge. The average doses were not significantly altered.
The fact that the physicians did not increase the prescription of ACEI/ARB and BB could have been due to the conclusion that symptom relief had been reached. Another plausible explanation could be fear of side effects. The 35% not receiving ACEI/ARB nor BB may have had low blood pressure, low heart rate or dizziness. The average systolic blood pressure among patients who received BB was 122, and among those who did not the average was 136 (p<0.010). There were no significant differences in blood pressure between patients who received ACEI/ARB and those who did not. The heart rate was on average 75 vs 78 and did not differ significantly. These data do not support a theory about considerations about hyper/hypotension or heart rate as reasons to refrain from treatment with BB or ACEI/ARB. Dizziness was not noted in the records.
The findings may also be related to the fact that further investigations to define HF type were not made, which suggests that there were no active considerations about type and level of HF, and hence a lower ambition to optimize treatment for each patient.
The increase of furosemide was not surprising, knowing that the patients were admitted due to acutely deteriorated HF, suffering from excess fluid, which makes this change probably clinically adequate. However, in the long run, the patient may have done better if adding or increasing the dose of ACEI/ARB would have been done instead of discharging patients with increased symptomatic diuretic treatment. Lee29 concluded that patients with the highest risk of death were less likely to receive ACEI or ARB.
A side effect of furosemide is the reduction in potassium30, leading to a need of substitution. This may have been one reason for the observed increased use of spironolactone, since a shift from furosemide to spironolactone would reduce both the need of furosemide and the need of potassium substitution31.
Subgroup with increased spironolactone
We identified a subgroup of patients who were either newly prescribed spironolactone during the care episode or had their spironolactone doses increased. This subgroup was found to have a greater weight loss during the care period, longer hospital stays, higher potassium levels, more NT-pro-BNP assessments and were more often referred to nursing homes. Although the doses of furosemide were not increased more in this group compared to the other patients, these patients had, due to the longer care stay, been treated with more furosemide, which could explain the greater weight-loss. Since diuretics induce lower potassium levels, a potassium sparing diuretic such as spironolactone could be added, allowing for a reduction of both furosemide and potassium salt, which can be seen as attractive to patients on high levels of diuretics.
Also, the choice of spironolactone may relate to the fact that the longer hospital stay gives the physician a few more days to evaluate an initiation of spironolactone treatment. The longer hospital stay may also have been related to waiting time for an available nursing home placement. We have previously shown that longer care episodes correlate to more prescription changes and higher quality in prescribing32. It may have been beneficial for these patients to spend some extra days in the geriatric clinic.
Among the parameters that we have studied, we cannot find explanations to why these patients were more often transferred to nursing homes. They did not differ from the other patients concerning age, gender, comorbidity burden, number of care episodes during the last year, EF, investigation status, or any other feature we have studied. However, data on frailty33 or any other functional score was not available.
Referral to primary care
The care chains depend on relevant information exchange when patients are transferred between caregivers. Kinugasa34 et al reported that the top three demands from general practitioners on information in the referrals concerning HF patients were: HF cause, BNP or NT-pro-BNP levels and data on EF or ECHO. Most patients in our study were discharged to their homes and referred to primary care for medical follow-up. We compared the information that was given in the referral to the geriatric clinic with the information that was written in the new referral to primary care after the hospital stay. We found that information about EF and etiology to a large extent was washed out and not transferred. Information about etiology was received from previous caregivers to the geriatric care for 31% of the patients and transferred from geriatric care to primary care for only 5%. The level of EF was received for 20% of the patients and transferred for 4%.
A reason for this wash-out-effect during the geriatric care episode may be the focus on fluid retention. Patients were to a very limited extent further investigated and therefore considerations about type and etiology of HF may have been overlooked. The focus on shortening of care episodes leaves less time for care and reflection. We have previously shown, that the length of the care episode was correlated to the number of drug changes and, in turn, to quality in drug prescribing32.
There may be a tradition in geriatric care to focus more on symptom relief than curing or disease modifying strategies. This could lead to less attention to etiology and treatment plans. The wash-out-effect could therefore be seen as a logical endpoint to a care episode. Further investigations were rarely performed and alterations in treatment performed only to a limited degree.