Ottawa Heart Failure Risk Scale (OHFRS)

Clinical background

Deciding whether to admit or discharge a patient with acute heart failure from the emergency department is one of the hardest disposition decisions in emergency medicine. A substantial proportion of patients are discharged from the emergency department after initial treatment, and some of them suffer serious events within days. At the same time there is no accepted, evidence-based algorithm for separating those who can be discharged from those who need a hospital bed. Admission rates vary widely between clinicians and countries, reflecting the lack of standardisation rather than differences in case mix.

The Ottawa Heart Failure Risk Scale (OHFRS) was developed to fill this gap: a structured, score-based tool that estimates the risk of a serious adverse event within 14 days and thereby supports the decision on admission or discharge for patients who are not already obviously high or low risk.

Calculating the Ottawa Heart Failure Risk Scale

The OHFRS is the sum of ten weighted binary variables, each assessed after 2 to 12 hours of treatment in the emergency department:

$$\text{OHFRS} = 1{\times}S_{\text{TIA}} + 2{\times}I_{\text{intub}} + 2{\times}\text{HR}$${\text{arrival}} + 1{\times}\text{SpO}2 + 1{\times}\text{HR}{\text{walk}} + 2{\times}\text{ECG}{\text{isch}} + 1{\times}U + 2{\times}\text{CO}_2 + 2{\times}T + 1{\times}N

where $S_{\text{TIA}}$ is a history of stroke or TIA, $I_{\text{intub}}$ is previous intubation for respiratory failure, $\text{HR}${\text{arrival}}HRarrival​ is a heart rate $\geq 110$/min on arrival, $\text{SpO}$2SpO2​ is an oxygen saturation $< 90$% on arrival, $\text{HR}${\text{walk}} is a heart rate $\geq 110$/min during a 3-minute walk test or an inability to perform it, $\text{ECG}${\text{isch}} is new ischaemic ECG changes, $U$ is a blood urea $> 12$ mmol/L, $\text{CO}_2$ is a serum CO2 (bicarbonate) $\geq 35$ mmol/L, $T$ is a troponin raised to infarction level and $N$ is an NT-proBNP $> 5000$ ng/L. The score ranges from 0 to 15.

The derivation cohort consisted of 559 adult patients with acute decompensated heart failure, enrolled prospectively at six large Canadian emergency departments between 2007 and 2010 [1]. The outcome measure was a composite serious adverse event (SAE): death, intubation, admission to a monitored unit, or relapse requiring admission. An SAE occurred in 11.6% of patients, almost half of them in those not initially admitted. The model was built with stepwise logistic regression and internally validated by bootstrapping over 1,000 replications [1].

Interpretation in practice

The OHFRS is constructed as a screening tool with high sensitivity: the instrument is designed not to miss patients at risk rather than to save hospital beds. The derivation study discussed thresholds of 1, 2 and 3 points for admission, with lower thresholds giving higher sensitivity but more admissions [1]. In the prospective validation study, an admission threshold of more than 1 point was used, which would have increased sensitivity markedly compared with actual clinical practice, but at the price of a higher admission rate [2].

Clinically, the score can be interpreted along the following lines:

An important addition is that NT-proBNP, when available, further improves the sensitivity of the instrument [2]. In the validation study, physicians expressed discomfort with using the OHFRS in only 11.9% of cases, which suggests acceptable clinical usability [2].

Validation and performance

The prospective validation was carried out at six tertiary hospital emergency departments with 1,100 patients (mean age 77.7 years) [2]. An SAE occurred in 15.5%, somewhat higher than in the derivation cohort. The validation confirmed that the OHFRS has high sensitivity for SAEs, but with a consequent increase in the admission rate. Where NT-proBNP values were available, sensitivity improved further [2].

A systematic review of risk scores for acute heart failure in the emergency department identified the OHFRS as one of eight scoring systems derived specifically in emergency departments with heart failure patients [3]. The review noted that the OHFRS had been used in two studies (derivation and validation), fewer than for competing instruments such as EHMRG (six studies) and MEESSI-AHF (twelve studies). The review concluded that two scoring systems, EHMRG and MEESSI-AHF, were sufficiently validated and accurate to be recommended for clinical decision support, while the OHFRS was not highlighted at that level [3].

A later analysis by the same research group pooled data from three prospective cohorts (2,246 patients in total at ten Canadian emergency departments) and found that a model with all 13 candidate variables had a c-statistic of 0.69 (95% CI: 0.66 to 0.72) for predicting the composite serious outcome [4]. This is modest discrimination, in line with what has been reported for other heart failure scores in the emergency department. The group also concluded that the original ten variables of the OHFRS might be too many for optimal use in a busy emergency department, and therefore developed a more concise successor, HEARTRISK6, with six variables [4].

Limitations

The OHFRS has several important limitations of which clinicians should be aware:

The requirement for a structured walk test. The instrument presupposes that a 3-minute walk test is carried out as part of the emergency assessment, after 2 to 12 hours of treatment. The patient is considered to have "failed" if the walk test cannot be started or completed because of abnormal vital signs. This requirement may be difficult to meet in a crowded and busy emergency department, and it presupposes trained staff.

Derived and validated only in Canada. Both the derivation and the validation cohorts consisted of Canadian patients at tertiary emergency departments. External validation in other health systems, with a different case mix and different admission practices, is lacking in the literature. The admission rate in the cohorts (38 to 49%) may differ markedly from local conditions.

Limited discrimination. With a c-statistic in the region of 0.69 for the full variable set in the pooled cohort [4], the ability of the instrument to separate high- from low-risk patients is modest. This is not unique to the OHFRS but reflects a general challenge for risk scores in acute heart failure.

Not recommended in international guidelines. The 2020 systematic review did not highlight the OHFRS among the scoring systems recommended for clinical use [3]. The 2016 ESC heart failure guidelines contained no specific recommendations on risk stratification in the emergency department [3].

Superseded by HEARTRISK6. The same research group that developed the OHFRS has since published HEARTRISK6, a more concise six-variable scale based on a larger pooled cohort [4]. HEARTRISK6 has not undergone independent external validation either, but it is designed to be more practicable in the clinic.

Excluded patients. The derivation study excluded patients who were obviously too unwell to consider for discharge (oxygen saturation $< 85$%, heart rate $\geq 120$/min, systolic blood pressure $< 85$ mmHg, confusion, acute ischaemic chest pain, STEMI, terminal illness, haemodialysis) [4]. The OHFRS therefore does not apply to these patients, in whom admission is indicated whatever the score.

References

1. Stiell IG, Clement CM, Brison RJ et al. A risk scoring system to identify emergency department patients with heart failure at high risk for serious adverse events. Academic Emergency Medicine 2013;20(1):17-26. PMID: 23570474
2. Stiell IG, Perry JJ, Clement CM et al. Prospective and Explicit Clinical Validation of the Ottawa Heart Failure Risk Scale, With and Without Use of Quantitative NT-proBNP. Academic Emergency Medicine 2017;24(3):316-327. PMID: 27976497
3. Miró Ò, Rossello X, Platz E et al. Risk stratification scores for patients with acute heart failure in the Emergency Department: A systematic review. European Heart Journal: Acute Cardiovascular Care 2020;9(5):375-398. PMID: 33191763
4. Stiell IG, Perry JJ, Eagles D et al. The HEARTRISK6 Scale: Predicting Short-Term Serious Outcomes in Emergency Department Acute Heart Failure Patients. JACC: Advances 2024;4(5):100988. PMID: 39129980