## Journal of the American College of Cardiology

# Doppler-derived dP/dt and −dP/dt predict survival in congestive heart failure

## Author + information

- Received January 14, 2000.
- Revision received April 25, 2000.
- Accepted June 26, 2000.
- Published online November 1, 2000.

## Author Information

- Theodore J Kolias, MD
^{a},* (tkolias{at}umich.edu), - Keith D Aaronson, MD
^{a}and - William F Armstrong, MD, FACC
^{a}

- ↵*Reprint requests and correspondence: Dr. Theodore J. Kolias, Division of Cardiology, University of Michigan Medical Center, L3119 Women’s, 1500 E Medical Center Dr, Ann Arbor, Michigan 48109-0273

## Abstract

OBJECTIVES

The purpose of this study was to evaluate the ability of novel Doppler indices of left ventricular (LV) systolic and diastolic function to predict survival in patients with congestive heart failure (CHF).

BACKGROUND

Congestive heart failure is associated with an increased risk of death or cardiac transplantation, yet techniques to predict survival are limited.

METHODS

Doppler-derived dP/dt and −dP/dt were determined prospectively from the continuous-wave Doppler spectrum of the mitral regurgitation jet (dP/dt = 32/time between 1 and 3 m/s; −dP/dt = 32/time between 3 and 1 m/s) in 56 patients with chronic CHF (age, 60 ± 15 years; LV ejection fraction, 23 ± 9%). Baseline clinical and echocardiographic variables were also obtained, and clinical follow-up was performed in all patients.

RESULTS

Twenty-four patients experienced a primary event of cardiac death (n = 15), United Network for Organ Sharing status I (inotrope-dependent) heart transplant (n = 3) or urgent implantation of a LV assist device (n = 6). Doppler-derived dP/dt (dichotomized to ≥ or <600 mm Hg/s; p = 0.0002) and −dP/dt (trichotomized to <450, 450 to 550 and >550 mm Hg/s; p = 0.0001) predicted event-free survival, as did Doppler-derived risk groups determined by the combination of the two (low risk, dP/dt ≥ 600; intermediate risk, dP/dt < 600 and −dP/dt ≥ 450; high risk, dP/dt < 600 and −dP/dt < 450; p = 0.0001). Multivariable analysis revealed Doppler-derived risk groups, intravenous inotrope requirement and blood urea nitrogen as significant independent predictors of outcome.

CONCLUSION

New Doppler indices of dP/dt, −dP/dt and risk groups defined by the combination of dP/dt and −dP/dt predict event-free survival in patients with CHF.

Severe congestive heart failure (CHF) is associated with a marked increase in risk of death or need for cardiac transplantation (1–3). The ability of current techniques to predict which patients will die or require heart transplantation is limited. Identification of such patients is important so that resources may be appropriately allocated and patients may understand their prognosis.

Previous investigators have evaluated markers of left ventricular (LV) systolic and diastolic function to predict outcome in heart failure. These markers have included ejection fraction (3–7) and transmitral inflow characteristics (8–12). One of the limitations of these markers, however, is their load-dependence (13,14), which may confound accurate assessment of LV systolic or diastolic function and limit their predictive value. By contrast, isovolumic phase indices of LV function, such as dP/dt and −dP/dt, are less load-dependent (13,15,16) and are theoretically a more accurate reflection of LV function. Their measure, however, has required invasive cardiac catheterization using a high fidelity micromanometer-tipped catheter, thus limiting their widespread use. Recently, a noninvasive echocardiographic method has been proposed to determine the isovolumic phase indices of cardiac function dP/dt (17–19) and −dP/dt (18–21).

Doppler-derived dP/dt and −dP/dt can be calculated from the continuous-wave Doppler spectrum of the mitral regurgitation (MR) jet obtained during transthoracic echocardiography. They have been shown to correlate with their corresponding invasively derived parameters (17,21). This method, however, has not been applied to heart failure patients to determine if it predicts survival. The present study was designed to test the hypothesis that Doppler-derived dP/dt and −dP/dt predict survival in patients with chronic CHF.

## Methods

Patients (n = 61) with chronic CHF and LV ejection fraction (LVEF) <50% were recruited for the study between March 1998 and November 1998. Five patients were excluded secondary to inadequate MR signals (n = 3) or mechanical valve prostheses (n = 2), leaving 56 patients who were enrolled. Each patient underwent a baseline history, physical examination and transthoracic echocardiogram, and clinical and echocardiographic variables were recorded. In addition, the baseline serum sodium, blood urea nitrogen and serum creatinine values were recorded.

The patients were then followed for the occurrence of primary events, which were defined as cardiac death, United Network for Organ Sharing (UNOS) status I heart transplantation or urgent implantation of a LV assist device. Patients not experiencing a primary event were censored at the time of UNOS status II heart transplant, noncardiac death or last follow-up. Patient follow-up was conducted by phone interviews and/or review of the hospital records and was complete in all patients. The study was approved by the institutional review board of the University of Michigan, and all patients provided informed consent.

### Echocardiographic examination

A transthoracic two-dimensional echocardiogram and Doppler examination were performed in all patients using an echocardiography system (Hewlett-Packard 2500, Hewlett-Packard 5500 [Hewlett-Packard; Andover, Massachusetts], Acuson Sequoia, or Acuson 128XP [Acuson Corporation; Mountain View, California]). The MR jet was interrogated with continuous-wave Doppler from an apical four-chamber or two-chamber view, using a sweep speed of 100 mm/s, and the baseline and scale were adjusted to maximize the spectral signal. Care was taken to align the imaging beam parallel to the direction of the regurgitant jet. Other echocardiographic parameters obtained included LV dimensions from the parasternal long axis views and the E/A ratio obtained using pulsed-wave Doppler from the apical four-chamber view with the sample volume placed at the mitral valve leaflet tips. The ejection fraction was calculated at baseline for each patient using the modified Simpson method from the apical four-chamber view.

### Doppler-derived dP/dt and −dP/dt measurements

Continuous-wave Doppler spectra of the MR jet were analyzed prospectively using an offline system (TomTec Imaging; Boulder, Colorado). Doppler-derived dP/dt was determined as follows: the two points on the MR spectrum corresponding to 1 m/s and 3 m/s were identified. These points corresponded to LV-left atrial pressure gradients of 4 mm Hg and 36 mm Hg using the modified Bernoulli equation (P = 4v^{2}). Doppler-derived dP/dt was defined as ΔP/Δt = 36-4/Δt = 32 mm Hg/Δt (Fig. 1). −dP/dt was determined from the diastolic slope of the MR spectrum, using the time required to go from 3 m/s to 1 m/s. The averages of three dP/dt measurements and three −dP/dt measurements were determined for each patient (five if they were in atrial fibrillation), thus yielding an averaged dP/dt and −dP/dt for each patient.

### Statistical analysis

Clinical and echocardiographic variables were evaluated as possible univariable predictors of survival by the Kaplan-Meier method and log-rank testing (categorical variables) or by univariable Cox proportional hazards models (continuous variables). Continuous variables were grouped into strata for Kaplan-Meier analysis. The proportional hazards assumption was confirmed graphically by log (−log survival) versus log time plots. Significant univariable predictors of outcome (p < 0.1) were then analyzed with multivariable Cox proportional hazards regression models. Statistical testing was two-tailed, and significance was defined as p < 0.05. Calculations were performed using statistical analysis software (SAS version 6.12).

## Results

The baseline clinical and echocardiographic characteristics of the patients in the study are summarized in Table 1. During follow-up, 24 (43%) of the patients experienced a primary event of cardiac death (n = 15), UNOS status I heart transplantation (n = 3) or urgent implantation of a LV assist device (n = 6). Of the 32 patients not experiencing a primary event, 29 were alive without transplant, 2 underwent UNOS status II heart transplantation and one developed lung cancer and died during follow-up.

Univariable analysis demonstrated that several clinical and echocardiographic variables significantly predicted the occurrence of a primary event during follow-up (Table 2). The strongest of these were intravenous inotrope requirement, Doppler-derived −dP/dt and Doppler-derived dP/dt. Of note, ejection fraction and E/A ratio >2.0 were not significant predictors of a primary event.

Evaluation of event-free survival for ranges of Doppler-derived dP/dt revealed two distinct strata: dP/dt < 600 and dP/dt ≥ 600. Patients with dP/dt < 600 mm Hg had worse event-free survival compared with those with dP/dt ≥ 600 (Fig. 2). Likewise, evaluation of event-free survival for ranges of Doppler-derived −dP/dt revealed three distinct strata: −dP/dt < 450, −dP/dt = 450–550 and −dP/dt > 550, with a significant overall difference among the groups (Fig. 3). The difference between −dP/dt < 450 and −dP/dt = 450–550 was also significant (p = 0.0428), as was the difference between −dP/dt < 450 and −dP/dt > 550 (p = 0.0001). The difference between −dP/dt = 450–550 and −dP/dt > 550, however, was not significant (p = 0.1357).

We then combined information from both dP/dt and −dP/dt to define three Doppler-derived risk groups based on the following criteria: low risk (dP/dt ≥ 600), intermediate risk (dP/dt < 600 plus −dP/dt ≥ 450) and high risk (dP/dt < 600 plus −dP/dt < 450). Survival analysis revealed both a significant overall difference as well as significant differences between the individual groups (Fig. 4). Using this risk stratification strategy, the 30-day event-free survival estimate of the high-risk group is 0.21 ± 0.11 (mean ± SD); by contrast, the 180-day survival estimates of the intermediate- and low-risk groups are 0.61 ± 0.15 and 0.81 ± 0.07, respectively. The sharp drop in survival of the low-risk group near the end of the study occurred because only a few patients remained at that point.

Multivariable proportional hazards regression modeling was performed using Doppler-derived risk groups together with other significant univariable clinical and echocardiographic predictors of outcome. This yielded only three significant independent predictors of outcome: Doppler-derived risk group, the requirement for intravenous inotropes, and blood urea nitrogen (Table 3). Using −dP/dt and dP/dt individually in place of Doppler-derived risk groups revealed −dP/dt as a significant independent predictor of outcome (p = 0.0004). Finally, if dP/dt was entered into the model in the absence of −dP/dt and Doppler-derived risk groups, it too was a significant independent predictor of outcome (p = 0.0373).

## Discussion

This study demonstrates the potential of Doppler-derived dP/dt and −dP/dt to predict survival in patients with chronic CHF. Low-, intermediate- and high-risk groups could be identified based on the combined parameters of Doppler-derived dP/dt and −dP/dt. The strong predictive value obtained by combining Doppler-derived dP/dt and −dP/dt illustrates the interplay of systolic and diastolic function in determining prognosis in patients with CHF. This has been suggested in previous studies in which impaired diastolic function as determined by mitral filling was a significant predictor of outcome in patients with low ejection fractions (8–10,12). The present study further supports the role of diastolic function in determining prognosis in patients with impaired systolic function and proposes a new sensitive index for combining parameters of systolic and diastolic function for the purpose of risk stratification.

Doppler-derived dP/dt was determined in this study by measuring the mean rate of pressure rise of the MR jet between 1 and 3 m/s, and as such it represents the mean dP/dt between 1 and 3 m/s. This method has been shown to correlate well with invasively derived peak dP/dt (17). Likewise, Doppler-derived −dP/dt was determined by measuring the mean rate of pressure fall of the MR jet between 3 and 1 m/s. The time required to go from 3 to 1 m/s has been shown previously to correlate reasonably well to the time constant of relaxation, tau (21). Although more sophisticated methods of determining dP/dt and tau using digitization of the MR velocity spectrum were previously found to correlate with invasive parameters more closely (18,21), we chose this simplified method for its ease of applicability and greater potential for widespread use.

Although these Doppler indices of cardiac function have been described previously, to our knowledge, they have never been applied to a CHF population to predict survival. Only a few clinical uses have been reported, such as the use of Doppler-derived dP/dt to predict postoperative ejection fraction (22) and inotropic requirement (23) in patients undergoing mitral valve surgery. Doppler-derived dP/dt and −dP/dt are well-suited, however, for application to patients with chronic CHF because most of these patients have some degree of central MR secondary to annular dilatation and chordal tethering. As seen in this study, obtaining dP/dt and −dP/dt is feasible in most patients. Furthermore, the noninvasive nature of this technique makes it easy to apply serially and in a variety of settings. In many cases, echocardiography and Doppler are routine aspects of management in patients with CHF, and obtaining these measures requires only minimal additional effort. Another advantage is their feasibility in patients with atrial fibrillation, in whom obtaining a mitral E/A ratio is not feasible.

Doppler-derived dP/dt and −dP/dt are sensitive measures of LV systolic and diastolic function, which may explain their enhanced ability to predict survival. Previous studies of other markers of LV systolic and diastolic function have shown that markers of worsening function are related to worse prognosis. Previously used markers include ejection fraction for systolic function and mitral inflow E/A ratio for diastolic function. In this study, Doppler-derived dP/dt and −dP/dt as well as the Doppler-derived risk groups outperformed both of these measures with respect to predicting survival. In addition, other previously described echocardiographic variables such as LV internal dimension at end-systole and LV internal dimension at end-diastole (24) were also less powerful than Doppler indices in predicting survival.

The validity of Doppler-derived dP/dt and −dt/dt as determined by taking the mean rate of pressure rise or fall between 1 to 3 or 3 to 1 m/s has been questioned in the past. The concern is that since these are measured during MR, they are not true isovolumic measures of function. Although a valid point, the change in atrial pressure during the time course of these measurements is small compared with the change in overall pressure, suggesting that its role is minor. In addition, our patient population included patients with all degrees of MR, suggesting that the applicability of these measures is widespread and not limited only to those patients with severe MR. Another concern with this technique is that it provides mean dP/dt rather than peak dP/dt as has been used in other studies. Since mean dP/dt is always calculated between 1 and 3 m/s, it is always determined at a constant developed pressure that is almost invariably prior to the opening of the aortic valve. In this regard, this technique may hold an advantage compared with peak dP/dt, since peak dP/dt often occurs near the time of opening of the aortic valve and may be more affected by afterload (25). Ultimately, the technique should be judged on its ability to contribute to the clinical management of the patient, and the present study suggests that the present method has potential to do so.

### Study limitations

A limitation of this study is that these were very ill patients: the majority of them were in New York Heart Association class III or IV, with several requiring intravenous inotropes. Whether these indices are able to predict outcome in a less ill, more ambulatory population of patients needs to be evaluated. In addition, we evaluated these indices in a relatively small group of patients. Further investigation in a larger population of heart failure patients is warranted. Finally, only one of the patients with a dP/dt ≥ 600 had a −dP/dt < 450; further study is warranted to determine if this subgroup has a different prognosis compared with patients with a dP/dt ≥ 600 combined with a −dP/dt ≥ 450.

## Conclusions

This study demonstrates that the Doppler indices of dP/dt and −dP/dt predict survival in patients with CHF. These simple, noninvasive measures have potential to play a role in the management of heart failure patients in the future.

## Acknowledgements

The authors gratefully acknowledge Mary Sue LeMire, Linda Dziekan, Diane Eberhart and Deborah Strong for their technical support.

- Abbreviations
- CHF
- congestive heart failure
- LV
- left ventricular
- LVEF
- left ventricular ejection fraction
- MR
- mitral regurgitation
- UNOS
- United Network for Organ Sharing

- Received January 14, 2000.
- Revision received April 25, 2000.
- Accepted June 26, 2000.

- American College of Cardiology

## References

- ↵
- ↵
- Aaronson K.D.,
- Schwartz J.S.,
- Chen T.M.,
- Wong K.L.,
- Goin J.E.,
- Mancini D.M.

- Gradman A.,
- Deedwania P.,
- Cody R.,
- et al.

- Johnson G.,
- Carson P.,
- Francis G.S.,
- Cohn J.N.

- ↵
- Xie G.Y.,
- Berk M.R.,
- Smith M.D.,
- Gurley J.C.,
- DeMaria A.N.

- Heart Muscle Disease Study Group,
- Pinamonti B.,
- Di Lenarda A.,
- Sinagra G.,
- Camerini F.

- Temporelli P.L.,
- Corra U.,
- Imparato A.,
- Bosimini E.,
- Scapellato F.,
- Giannuzzi P.

- ↵
- Kass D.A.,
- Maughan W.L.,
- Guo Z.M.,
- Kono A.,
- Sunagawa K.,
- Sagawa K.

- Choong C.Y.,
- Herrmann H.C.,
- Weyman A.E.,
- Fifer M.A.

- Quinones M.A.,
- Gaasch W.H.,
- Alexander J.K.

- Starling M.R.,
- Montgomery D.G.,
- Mancini G.B.J.,
- Walsh R.A.

- ↵
- Bargiggia G.S.,
- Bertucci C.,
- Recusani F.,
- et al.

- ↵
- Chen C.,
- Rodriguez L.,
- Guerrero L.,
- et al.

- Chen C.,
- Rodriguez L.,
- Lethor J.P.,
- et al.

- Chen C.,
- Rodriguez L.,
- Levine R.A.,
- Weyman A.E.,
- Thomas J.D.

- ↵
- Nishimura R.A.,
- Schwartz R.S.,
- Tajik A.J.,
- Holmes D.R.

- ↵
- Pai R.G.,
- Bansal R.C.,
- Shah P.M.

- ↵
- ↵
- ↵
- Mason D.T.,
- Braunwald E.,
- Covell J.W.,
- Sonnenblick E.H.,
- Ross J. Jr.