Background
Heart failure (HF) is common and its impact on a patient’s quality of life (QoL) depends on its severity [
1]. Finding simple, cheap, and effective ways to evaluate the severity of HF is important. The severity of HF is usually classified according to the patient’s medical history and perceived effort tolerance, which forms the basis of the New York Heart Association (NYHA) functional classification (FC) of HF. Affective interference in the part of patients and cognitive impressions formed by physicians render this approach fairly subjective, especially with regard to NYHA FC II and NYHA FC III. Although a more precise classification based on cardiopulmonary exercise tests (CPET) was proposed by Fu et al. in 2011 [
2], this method may be clinically impractical, because of feasibility and cost considerations. Furthermore, assessing a patient’s QoL using questionnaires, including the Short Form (SF)-36, is time-consuming in routine clinical practice because of the large amounts of time required for their completion and the subsequent complex analyses. A simple and cheap “bedside” tool that provides a composite measure of the symptoms and their effects on a patient’s QoL would be useful to describe the severity of the clinical syndrome in individual patients.
Physicians who practice traditional Chinese medicine (TCM) use four diagnostic approaches, i.e., inspection, auscultation and olfaction, interrogation, and palpation to record patients’ information in relation to specific patterns [
3,
4]. TCM offers a simple, cheap, and non-invasive approach for evaluation of the severity of HF and does not involve use of expensive equipment. There are several specific patterns associated with the clinical presentation of HF from the perspective of a comprehensive approach that uses TCM [
5‐
9]. Irrespective of these patterns, the main manifestations of HF include palpitations, dyspnea, and fatigue [
10,
11]. “Panting on exertion” is used to grade the severity of HF according to NYHA FC [
10,
12]. An integrated approach comprising the SF-36, Minnesota Living with Heart Failure Questionnaire (MLHFQ), and a TCM model, which incorporates a TCM inquiry list, could provide another method for evaluating the severity of HF. Furthermore, this integrated method could identify critical parameters that may differentiate between NYHA FC II and NYHA FC III.
The aim of this study was to design an inquiry list from the perspective of TCM (called the TCM inquiry list) that could be used in routine clinical practice to resolve the current problems associated with the FC of HF.
Discussion
The ambiguity of the criteria used to determine an HF patient’s classification as NYHA FC II or NYHA FC III and the large economic burden associated with the exercise tests used to distinguish between these classifications prompted the development of a TCM inquiry list that could be used in routine clinical practice. The findings of this study demonstrate that the TCM inquiry list for HF correlated well with the mental and physical aspects of the QoL assessed using the SF-36 and MLHFQ. When combined with the MLHFQ, this novel TCM inquiry list had a correct prediction rate of 70.4 % and could distinguish between NYHA FC II and NYHA FC III.
The five factors extracted from the TCM inquiry list using exploratory factor analysis were Qi Depression (factor 1), Heart Qi Vacuity and Blood Stasis (factor 2), Heart Blood Vacuity (factor 3), Dual Qi-Blood Vacuity (factor 4), and Yang Vacuity (factor 5), as shown in Table
4. These factors are partially compatible with the patterns associated with the clinical perspective of TCM in the HF population, namely, Heart-Lung Qi Vacuity, Dual Vacuity of Qi and Yin, Heart-Kidney Yang Vacuity, and Qi Vacuity with Blood Stasis [
11]. However, Yang Vacuity Water Flood, Phlegm-Damp Obstructing the Lung, and Exhaustion of Yin and Desertion of Yang did not emerge in the current analysis, and these factors may result in selection bias [
11]. Only factors 1–4 showed significant differences between NYHA FC II and NYHA FC III (Table
7), and can, therefore, be used to distinguish between NYHA FC II and NYHA FC III. Factor 5, which hints at Yang Vacuity, should be present in patients with more severe HF, for example, NYHA FC IV [
7]. The HF patients in this study were classified into NYHA FC II and NYHA FC III; hence, factor 5 could not show a significant difference between these groups. The patterns that describe Yang Vacuity Water Flood, Phlegm-Damp Obstructing the Lung, and Exhaustion of Yin and Desertion of Yang did not emerge in the present analysis, and these patterns may also be present in patients with severe HF [
5‐
9]. Without any prediction or interference, the factor analysis essentially extracted these five factors, which are partially compatible with the current TCM pattern (or “Zheng”). This result could in turn confirm the existence of a TCM pattern.
Table 7
The difference of the factor score between Fc II and III
TCM inquiry | 50.6 ± 29.6 | 76.0 ± 53.2* | 0.59 | 0.87 |
Factoc 1 | 11.7 ± 9.8 | 18.7 ± 21.2* | 0.42 | 0.72 |
Factoc 2 | 12.3 ± 9.0 | 18.9 ± 15.1* | 0.53 | 0.82 |
Factoc 3 | 7.9 ± 5.5 | 11.9 ± 9.9* | 0.50 | 0.80 |
Factoc 4 | 6.1 ± 5.1 | 10.2 ± 8.6* | 0.58 | 0.86 |
Factoc 5 | 8.7 ± 9.1 | 11.7 ± 10.3 | 0.31 | 0.37 |
When the NYHA FC II and NYHA FC III categories were used as binary variables and the five TCM factors were entered into the logistic regression model to predict NYHA FC, only factor 4 was significantly associated with differences between NYHA FC II and NYHA FC III. Hence, factor 4 is the critical parameter from the perspective of TCM that can help to differentiate between NYHA FC II and NYHA FC III in a population with HF.
The typical clinical manifestations of HF include dyspnea on exertion, nocturnal dyspnea, and leg swelling, and inquiries about these manifestations often comprise parts of the current questionnaires. Factor 4, which includes Q1 (dizziness), Q2 (dizzy vision), and Q11 (general fatigue), and hints at Dual Qi-Blood Vacuity, which is the key pattern associated with coronary artery disease in TCM [
7,
8], comprises distinct themes within TCM that are derived from the typical clinical manifestations of HF. Moreover, these three items are the key elements that are associated with the risk of falls [
16,
17]. Hence, we considered that integrating this factor from the TCM inquiry list into other questionnaires may increase the prediction rate of HF severity. Since the SF-36 is a structured questionnaire with a specific scoring system, we chose the MLHFQ for integration, and when the PCS from the MLHFQ and the score from factor 4 on the TCM inquiry list were combined, the correct prediction rate rose to 70.4 %, as expected.
Table
8 presents the associations between the score from factor 4 on the TCM inquiry list and the physiologic parameters obtained during the exercise tests, which extracted the latent physiologic traits associated with factor 4. The factor 4 score had a significant negative correlation with peak VO
2, peak workload, VO
2 at the anaerobic threshold, oxygenated hemoglobin levels in muscle at the anaerobic threshold, total cerebral hemoglobin level at peak exercise, and the oxygen uptake efficiency slope, and very weak but significant positive correlations with cerebral and muscle deoxygenated hemoglobin levels during peak exercise. The deoxygenated state of the brain and the low cerebral blood flow during the peak exercise phase may be associated with dizziness and dizzy vision. The deoxygenated state of the muscles during the peak exercise phase may be associated with limb weakness. This indicates that the differences between NYHA FC II and NYHA FC III are associated with the blood supply to the brain and muscles, which is in agreement with previous studies [
2,
18]. Cardiologists rarely ask patients about dizziness and vertigo when they are assessing the severity of HF, so these items could be integrated into future clinical evaluations of HF severity.
Table 8
Pearson’s coefficient of factor 4 score with physiologic parameters
Peak oxygen consumption | -.266(0.004)** |
Peak cardiac output | -.065(0.490) |
LVEF | -.067(0.478) |
Peak_workload | -.244(0.008)** |
VO2 at AT level | -.196(0.047)* |
AT_C_O2Hb | -.040(0.690) |
AT_C_HHb | -.017(0.864) |
AT_C_THb | -.040(0.685) |
AT_M_O2Hb | -.224(0.023)* |
AT_M_HHb | .004(0.967) |
AT_M_THb | -.092(0.357) |
Peak_C_O2Hb | -.229(0.014)* |
Peak_C_HHb | -.230(0.013)* |
Peak_C_THb | -.195(0.037)* |
Peak_M_O2Hb | -.061(0.518) |
Peak_M_HHb | -.184(0.049)* |
Peak_M_THb | -.096(0.308) |
VE-VCO2 slope | .131(0.164) |
OUES | -.213(0.022)* |
This study has several limitations. Our HF population had similar PCS scores, but lower MLHFQ and worse MCS scores than those reported by other researchers [
19,
20]. Weaker associations between peak VO
2, the MLHFQ score, and the MCS were also noted. Poor correlations between MLHFQ score and CPET parameters have been described in a previous study [
21]; however, the correlations between the PCS, MCS, and MLHFQ (Table
3) were similar to those determined in other studies [
19,
20,
22]. Possible explanations for this phenomenon are as follows. First, the populations in previous studies were younger and novel NYHA FC criteria were used, namely, the aerobic capacity derived from CPET. Second, the status of patients recruited for the current study was not consistent, and some had undergone exercise training, which may have been a source of bias when the data were analyzed [
23,
24].
Q13 was removed from the factor analysis. That means that there is room for improvement in the choice of items included in this inquiry list. The Q13 item was removed from the exploratory factor analysis because of insufficient loading (<0.5 or less) in all factors, or excessive loading (>0.5 or more) in an excessive number of factors (>2 or more). For our purposes, this implies that either Q13 is not an important or common item or that there are surrogate items for Q13 that can be used instead. Therefore, subjects may choose other items first. Deletion of Q13 in this study did not change the final result. Hence, it was reasonable to delete this item.
The correlation coefficients demonstrated relationships between the factors, aerobic capacity, and QoL. We found that the factors in the TCM inquiry list had moderate correlations with the QoL questionnaires, but only weak correlations or no correlations at all with aerobic capacity. This may mean that aerobic capacity did not actually correlate with the TCM inquiry list and the QoL; hence, a non-linear relationship may have existed between QoL and the symptoms associated with the TCM inquiry list. Therefore, further analysis of the data is needed. In addition, the sample size in this study was barely adequate for development of a questionnaire, but should not be considered unsatisfactory [
15]. Moreover, we integrated the MLHFQ and the TCM inquiry list scores, because calculating the SF-36 score requires several mathematical procedures, and the total MLHFQ score only required the sum to be calculated. Finally, use of this novel TCM inquiry list to predict NYHA FC was associated with acceptable sensitivity, but the specificity, or false-positivity, was unsatisfactory. In clinical practice, high false positivity leads to a waste of medical resources, while elevated false negativity increases the medical risk. However, sacrificing specificity is reasonable in the high-risk practice of HF management. Further refinements would be necessary before implementing this TCM inquiry list. First, the questions for each item should be designed in such a way that they reflect the status of the patient. For instance, the question we designed for panting on exertion, i.e., “short of breath after walking on the flat road for 5 min” is not adequate to distinguish between NYHA Fc II and NYHA Fc III. “Short of breath on climbing two flights of stairs” may be more appropriate. Second, the patient’s response should be checked by an experienced health care professional to confirm the response is accurate [
25]. Third, including three of the other TCM diagnostic tools, i.e., inspection, palpation, and examination by listening and smell, could enable a more comprehensive evaluation.
Funding
This investigation was supported by grant from Chang Gung Medical Foundation Chang Gung Memorial Hospital, Keelung (CMRPG2D0231) and China Medical University under the Aim for Top University Plan of the Ministry of Education, Taiwan.