Introduction

Heart failure (HF) is one of the most common leading causes of death, and it also has high morbidity, which is a high burden to the health care system.1 Because of limited resources, risk stratification is necessary in HF patients, which will not only allow us to differentiate between high-risk and low-risk patients but also select appropriate treatment. There are several scores based on multiple parameters now available to predict mortality and cardiovascular events in patients with HF. One of the scores is get with the guidelines (GWTG) risk score, which was established by Peterson et al.2 It is basically established from a cohort study which has a role in predicting the prognosis of acute HF and also in chronic phase after discharge. There is sparse data available, especially from south India regarding utility of this GWTG score. So, our aim is to see the applicability of predicted GWTG risk score of HF among south Indian heart failure patients admitted to the intensive cardiac care (ICCU) of our hospital.

Aim

Our primary aim is to see the applicability of predicted GWTG risk score of HF among south Indian heart failure patients admitted to the ICCU of our hospital. The secondary aim is to see the event rates and correlate predicted GWTG risk score of heart failure with in-hospital complications.

Materials and Methods

We included all the patients admitted to the ICCU with the diagnosis of either ischemic or dilated cardiomyopathy (ICMP or DCMP) over 6 months (January 2018 to June 2018). Indication for admission was either symptomatic HF or to evaluate cause for HF. Patients of HF due to organic valvular heart diseases or due to pulmonary arterial hypertension (primary or secondary) were excluded. We recorded the demographic and clinical parameters along with the ECG, 2D echo features, and relevant laboratory investigations. In-hospital complications like recurrence or worsening of HF, requirement of ventilation, later development of cardiogenic shock or pericardial effusion, and death were noted.

Results

A total of 130 patients with HF were admitted to the ICCU of our unit during the study period, of which 97 patients fulfilled the inclusion criteria. These 97 patients had either ICMP or DCMP. Out of them, 65 were males (Fig. 1), with most of the patients between 40 to 80 years of age (Table 1).

Close

Fig. 1 Pie diagram showing sex distribution in study population.

Export to PPT

Fig. 1 Pie diagram showing sex distribution in study population.

The history of the patients showed half the patients were diabetic and had abnormal electrocardiogram (ECG), and more than half were hypertensive, had clinically left ventricular failure (LVF) and diagnosed with ICMP. Very few were smokers, and < 30% were alcohol consumers, had abnormal liver function tests (LFT) and diagnosed with DCMP (Table 2). In Table 3, clinical and laboratory parameters of the patients were mentioned.

The population was divided into ICMP patients and DCMP patients and all the variables were compared (Table 4). The average of systolic BP (SBP), pulse and the mean of GWTG score were greater in 22 patients afflicted with DCMP. There was no difference in age and ejection fraction (EF) between the two groups. As expected, coronary artery disease (CAD) risk factors were more frequent in ICMP patients who were statistically significant. DCMP patients were more tachycardic with low SBP, abnormal ECG, and more GWTG score. However, abnormal renal functions with anemia and more NT-pro-brain natriuretic peptide (NT-proBNP) elevations were observed in ICMP HF patients.

Concomitant diseases were mentioned in Table 5. Cardiogenic shock (SBP < 90 mm of Hg), complete heart block (CHB), chronic obstructive pulmonary diseases (COPD), and chronic kidney disease (CKD) were more frequent in ICMP group, whereas atrial fibrillation (AF) and infection precipitating congestive heart failure (CCF) were more with DCMP patients.

GWTG HF risk score helps to predict in-hospital all-cause HF mortality. Major clinical data was used to get the score numbers. These score numbers were divided into groups, and the total population of 97 were placed into these groups as per their scores (Table 6). As much as 70% of the population was in group 2. There were no patients with score > 58 in our study.

Out of 97 HF patients, 18 patients showed in-hospital complications. The details of complication are mentioned in Table 7. Out of six ventilated patients, five patients died. Even though pericardial effusion may be due to CCF, but it fell under the purview of in-hospital complications, as this was not present, and at the time of admission, it appeared later in the hospital course. In all cases, it was mild effusion only and did not contribute to hemodynamic compromise. Eight patients showed initial stabilization, were shifted to the wards, and required transfer to ICCU due to recurrence of HF in hospital. Cardiogenic shock was there in 8 patients, and 4 patients were in refractory shock even with maximum inotropic support after 48 hours, but none developed shock requiring inotrope support after the admission.

In Table 8, details of five died patients were given. All five patients admitted experienced cardiogenic shock with tachycardia and severe LV dysfunction (ejection fraction [EF] < 15%), renal failure, hyponatremia, NT-proBNP levels > 25,000 pg/dl and hepatic derangement.

The important prediction of GWTG risk score is the mortality percentage for hospitalized HF patients. Our primary aim was to correlate the GETG score with actual mortality in our study patients. From the total population of 97 HF patients, five (5%) deaths occurred with 45.00 mean GWTG and 92 discharged with 39.022 mean GWTG score (Fig. 2).

Close

Fig. 2 Box and interval plot of the get with the guidelines (GWTG) score of died and discharged patients.

Export to PPT

Fig. 2 Box and interval plot of the get with the guidelines (GWTG) score of died and discharged patients.

From the boxplot, we can say the Q1 of the GWTG score of the death population was 39.00 and Q3 was 50.50, whereas Q1 of the GWTG score of the discharge patients was 34.00 and Q3 was 44.00. The interval plot of the GWTG score has no overlapping, so we can say that there was a significant difference between the mean GWTG scores of patients in their final status (discharged/death) (p = 0.040).

Our secondary aim was to compare the GWTG score with in-hospital events even though this score is used only for in-hospital mortality. In Fig. 3, GWTG scores were compared between patients with in-hospital complication versus those without.

Close

Fig. 3 Box and interval plot of get with the guidelines (GWTG) score of patients with or without in-hospital complications.

Export to PPT

Fig. 3 Box and interval plot of get with the guidelines (GWTG) score of patients with or without in-hospital complications.

The boxplot of the GWTG score of the 79 patients without in-hospital complications showed a mean score of 38.038, with first quartile (Q1) 33.00 and third quartile 42.00. And the boxplot of the GWTG scores of the 18 patients with in-hospital complications show the mean score of 45.00, Q1 39.50 and Q3 51.00. The GWTG score was different in patients with or without in-hospital complications (38 vs. 45), which was statistically significant (p = 0.000). The interference of both the population GWTG scores were different, as patients with in-hospital complications had higher GWTG values (on an average 7 scores higher) than the GWTG score of the patients without (p = 0.000).

Discussion

American Heart Association (AHA) GWTG risk score predicts the mortality in hospital admitted patients with both reduced and preserved LV systolic function. In these patients, GWTG risk score usage is different, appropriate, and easy when compared with other risk stratification methods. The seven variables which are included in GWTG score are routinely collected at the time of admission4 Heart rate, nonBlack race, COPD, and serum sodium are independent predictors of mortality, but they were less important with regard to the points contributed by them to the total GWTG HF risk score. LV systolic function which is a risk predictor in some models is not present in this risk score3. So, absence of LVEF at time of admission does not limit prognostication; hence, using this model and this risk score has identical characteristics operating in both group of patients, so that this score can be applicable to a broad spectrum of heart failure patients.

As per standard guidelines, we have calculated GWTG score from race, age, SBP, heart rate, blood urea nitrogen (BUN) level, sodium levels, and presence of COPD. Presence of Black race (points 0 or 3), COPD (0 or 2), age between 19 to 110 years (0–28), SBP 50 to 200 (0–28), BUN 9 to 150 (0–28), heart rate 79 to 105 (0–8), and serum sodium 130 to 139 (0–4) scoring is given. Then, we have calculated GWTG scores. Older age, low SBP, elevated heart rate, low serum sodium, elevated BUN, presence of COPD, and nonBlack race are predicted to have increased risk of death.4 Age, SBP, and BUN contributed more to the overall point score, whereas heart rate, presence of COPD, serum sodium, and nonBlack race contributed relatively few points to the overall score. Subsequently, patients were divided as groups from 1 to 9, according to the GWTG score (0–79). In our study, most of the patients come under group 2 (70%, N = 68) with GWTG score of 34 to 50 (Table 9).

There are various risk scoring systems available to predict in-hospital mortality, but these available risk stratification models have some limitations.

The Acute Decompensated Heart Failure National Registry (ADHERE) study identified serum creatinine, BUN and SBP as the best predictors of in-hospital mortality using classification and regression tree analysis.5 As it uses only three variables, it is appealing to classify patients as low, intermediate, or high-risk. However, it does not include all variables which give more precise characterization of individual risk and significantly inform outcomes. The limitation is it gives overly high mortality of the low-risk group in comparison with other models. Likewise, acute heart failure (AHF) index has same limit.

OPTIMIZE-HF include variables like race, serum creatinine, serum sodium, age, HR, liver disease, previous cerebrovascular accident (CVA)/transient ischemic attack (TIA), peripheral vascular disease, left ventricular systolic dysfunction, COPD, SBP, previous HF hospitalization and limits, in that it does not include data on BUN at admission and does not have a separate derivation and validation cohort.6

The international “placebo-controlled randomized study of the selective A1 adenosine receptor antagonist rolofylline for patients hospitalized with acute decompensated heart failure and volume overload to assess treatment effect on congestion and renal function” (PROTECT) trial enrolled 2033 patients hospitalized with AHF and with mild-to-moderate renal function impairment. The limitation of this trial is that it did not include patients taking inotropic agents, patients with preserved EF, recent MI, and those with severe pulmonary disease; therefore, so the fact that it cannot be applicable to a wide range of community-based populations is its limitation.7

Enhanced Feedback for Effective Cardiac Treatment (EFFECT) and Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbations of Chronic Heart Failure (OPTIME-CHF) models. However, both these trials were established for 30- and 60-days mortality after hospitalization, and the OPTIME-CHF trial is not applicable to broad spectrum of patients, as it is limited to patients cohorts with LV systolic dysfunction only. Additional predictive models for long-term mortality have been reported but limited by the use of large numbers of variables, derivation in clinical trial populations, and inclusion of variables that are not typically assessed in clinical care, such as uric acid, peak oxygen consumption, or timing of diagnosis of HF.8 9

GWTG- HF risk score has several advantages and strengths and it differs from other risk scores, as it is based on patients with diverse demographic variables and wide range of comorbidities. With regard to LV systolic function, as it is applicable for HF with both reduced and preserved EF, it can be widely applicable. Additionally, the variables which were used are small in number and can be collected at the time of admission. Several other models of risk score prediction have more than 20 variables.10 They are not feasible and difficult to use in routine clinical practice as compared with GWTG risk score.

Hauptman et al11 found that among physicians caring for patients with HF, only a few clinicians can predict death accurately. Generally, physicians substantially overestimate the risk and overutilize the resources in critical care. Conversely, it demonstrates that medical therapy for HF is poorly calibrated to risk, with high-risk patients paradoxically receiving evidence-based medical therapy less often.12

Our study results are comparable with previous studies in case of in-hospital mortality. High GWTG score is associated with a greater death rate. Our study revealed that high GWTG score is also associated with a high incidence of in-hospital events.

Thus GWTG HF risk score can be used for patient risk quantification, thus facilitating patient triage and encouraging the use of evidence-based therapy in high-risk patients, so that therapy can efficiently be used for high-risk patients by reducing utilization in low-risk patients.