Dig Dis Sci DOI 10.1007/s10620-014-3192-5
ORIGINAL ARTICLE
Role of Heart Rate Variability in Predicting the Severity of Severe Acute Pancreatitis Luyao Zhang • Jing Zhou • Lu Ke • Yao Nie • Zhihui Tong • Weiqin Li • Jieshou Li
Received: 21 August 2013 / Accepted: 28 April 2014 Ó Springer Science+Business Media New York 2014
Abstract Background Infected pancreatic necrosis (IPN) and multiple organ dysfunction syndrome (MODS) are major complications of acute pancreatitis which determine disease severity and outcome. Aims The aim of this study is to investigate the value of admission heart rate variability as a marker of IPN or MODS in severe acute pancreatitis (SAP) patients. Methods Forty-one SAP patients within 72 h of symptoms onset were included in this prospective observational study. General demographics, laboratory data and the acute physiology and chronic health evaluation (APACHE) II scores were recorded at admission. 5-minute ECG signals were obtained at the same time for heart rate variability analyses to assess SAP severity.
L. Zhang J. Zhou L. Ke Y. Nie Z. Tong W. Li (&) J. Li Research Institute of General Surgery, Jinling Hospital, Medical School of Nanjing University, 305 East Zhongshan Road, Nanjing 210002, Jiangsu Province, China e-mail: [email protected] L. Zhang e-mail: [email protected] J. Zhou e-mail: [email protected] L. Ke e-mail: [email protected] Y. Nie e-mail: [email protected] Z. Tong e-mail: [email protected] J. Li e-mail: [email protected]
Results The baseline heart rate variability measurements, levels of low frequency/high frequency (LF/HF) were significantly lower whereas high frequency norm (nHF) levels were significantly higher in patients who present with IPN and MODS or died (P \ 0.01). Low frequency (LF) levels were lower in patients who present with IPN or MODS as compared to patients without these complications. Levels of low frequency norm (nLF) were lower in MODS and nonsurvival patients. nHF and LF/HF were good predictors of IPN and MODS, superior to procalcitonin. nHF and LF/HF were better than APACHE II in predicting IPN and LF/HF showed superiority over APACHE II in the prediction of MODS. Conclusions Admission heart rate variability is a good marker of IPN and MODS in SAP patients. Keywords Heart rate variability Infected pancreatic necrosis Multiple organ dysfunction syndrome Severe acute pancreatitis
Introduction Acute pancreatitis is the inflammation of pancreas and usually takes a self-limiting course. However, about 10–20 % acute pancreatitis patients progress to SAP, which is characterized by intensive inflammatory response in the early phase and IPN in the later phase [1]. SAP patients accompanying infection of pancreatic necrosis or organ failure are at high risk of mortality. The mortality increases to approximately 30 % when IPN is present and could reach as high as 46 % when MODS is present [2, 3]. Therefore, early prediction and diagnosis of IPN and organ failure are crucial in determinations of further treatment and prognosis.
123
Dig Dis Sci
Autonomic dysfunction occurs in many inflammatory diseases, including rheumatoid arthritis, inflammatory bowel disease, sepsis, and has been proved to be correlated with their morbidity and mortality [4–7]. Autonomic function could be well evaluated through three different methods: heart rate variability, baroreflex sensitivity and chemoreflex sensitivity. Heart rate variability is an indirect estimator of autonomic modulation of cardiovascular system and most commonly used in clinical research. Studies have suggested heart rate variability analysis providing early diagnosis and successful prognostication of infection in critically ill patients [8]. More specifically, decreased LF variation, a sign of impaired sympathetic activity, has been shown to closely associate with presence and severity of systemic infection [8]. Besides, heart rate variability could also predict prognosis in critical ill patients. Julio Pontet found that a reduction of heart rate variability could precede onset of MODS in septic shock patients [9]. In healthy people, the power of LF/HF ascends along with sympathetic activities. LF/HF provides an index of the dynamics of cardiac sympatho-vagal balance. Increase in parasympathetic tone over sympathetic tone (decrease in LF/HF) was identified as powerful predictor of mortality in cardiac risk patients undergoing major non-cardiac surgery and neurosurgical critically ill patients [8]. However, scarce attention has been paid to autonomic function in SAP patients. In the present study, we aimed to investigate the possible association between heart rate variability and disease severity in the acute phase of SAP. Furthermore, we also attempted to see whether heart rate variability can predict two determinant factors of severity in SAP: infection of pancreatic necrosis and organ failure.
Materials and Methods Patients A total of 41 adult patients with a diagnosis of SAP admitted at surgical intensive care unit of Jinling hospital within 72 h of abdominal pain onset from January 2012 to January 2013 were consecutively enrolled in this study. Diagnostic criteria of SAP were defined according to the Atlanta criteria [10]. Patients who presented one or more of the following conditions were excluded: (1) chronic pancreatitis, (2) pre-existing organ failure, (3) received surgical treatment before admission, (4) persistent arrhythmia and (5) use of cardiac pacemaker. The study protocol was approved by the ethics committee of Jinling hospital, and informed consent was obtained from the patient or his/her legal guardian. All patients received standard treatment for SAP [11].
123
Data Collection Demographic information and clinical relevant data of the patients were recorded. The APACHE II [12]. and SOFA (sequential organ failure assessment) scores [13] were calculated within 24 h on ICU admission. According to the new definition brought up by Dellinger et al. [14], IPN was defined when at least one of the following was present: (1) gas bubbles within (peri)pancreatic necrosis on computed tomography, (2) a positive culture of (peri)pancreatic necrosis obtained by image-guided fine-needle aspiration and (3) a positive culture of (peri)pancreatic necrosis obtained during the first drainage and/or necrosectomy. Organ failure was confirmed in three organ systems (cardiovascular, respiratory and renal) when a score of two or more according to the SOFA score presented [15]. MODS was defined as at least two of these three organ failures existed. In all patients, plasma procalcitonin (PCT) levels were measured on admission by chemoluminescent immunoassay (LUMITEST-PCT, BRAHMS Diagnostica AG, Hennigsdorf, Germany) in our central laboratory. Heart Rate Variability Analysis Measurement of heart rate variability was performed within 24 h of ICU admission. A 5-min ECG was obtained at bedside between 9:00 and 11:00 am using a 12-lead electrocardiogram machine (EDAN, Shenzhen, China). Heart rate variability was analyzed according to the standards developed by the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology [16]. Normal beat intervals were identified from the lead II ECG recordings by the Matlab software (MathWorks, Natick, MA). All artifacts and ectopic beats were removed in accordance with the guidelines outlined by the Taskforce of the European Society of Cardiology. The results were analyzed with both ‘‘time domain’’ and ‘‘frequency domain’’ methodologies [17, 18]. We calculated SDNN, RMSSD in the time domain. RMSSD primarily reflected vagus nerve-mediated changes in heart rate. SDNN was a marker of overall heart rate variability. The power spectrum of the beat-to-beat intervals was generated using fast Fourier transformation. The indexes of frequency domain calculated included: (a) VLF, in the frequency range less than 0.04 Hz and a dubious measure which did not have a specific physiological meaning. (b) LF, in the range of 0.04–0.15 Hz, correlated with sympathetically and vagally mediated control. (c) HF, in the region of 0.15–0.4 Hz, reflected vagally mediated control of heart period. (d) the LF/HF ratio, was used as the index of sympatho-vagal balance. HF and LF were also measured in their normalized forms, which represented the relative value of
Dig Dis Sci Table 1 Time domain and frequency domain measures of heart rate variability
Table 2 Characteristic of 41 SAP patients Age(yeas)
47 (34–59)
Variable
Units
Description
Gender(male/female)
22/19
VLF
ms2
Power in very low frequency range (B0.04 Hz)
Etiology, n (%)
LF nLF
ms
2
n.u. 2
HF
ms
nHF
n.u.
LF/HF SDNN RMSSP
ms ms
Power in low frequency range (0.04–0.15 Hz)
Gallstones
20 (50 %)
LF power in normalized units
Alcohol
7 (17 %)
LF/(Total Power–VLF) 9 100
Hyperlipidemia
9 (22 %)
Power in high frequency range (0.15–0.4 Hz)
Others
5 (12 %)
HF power in normalized units
APACHE II score
14 (11–18)
HF/(Total Power–VLF) 9 100 Ratio LF (ms2)/HF (ms2)
Ranson score
5 (4–7)
PCT level on admission(ng/mL)
0.3 (0.1–1.2)
Standard deviation of normal RR intervals
Balthazar index
6 (4–7)
Square root of the mean of the sum of the squares of differences between adjacent NN intervals
MODS
9 (22.0 %)
Sepsis
2 (4.9 %)
Pressure support
5 (12 %)
Surgical intervention IPN
15 (37 %) 16 (39 %)
Mortality
4 (9.8 %)
Length of ICU stay
9 (3–11)
Length of hospital stay
17 (8–18)
each power component in proportion to the total power minus the VLF component. The nHF reflected the modulation of vagus nerve discharge caused by respiration, nLF was the expression of baroreceptor-mediated regulation and due to the contribution of parasympathetic and mainly sympathetic discharge. The description of indexes we evaluated were set out in Table 1. Statistical Analysis Data were presented with its median and interquartile range (25–75th percentiles). Descriptive data were described as a absolute numbers or a percentage number. Mann–Whitney U test was used to compare continuous variables. The difference of heart rate variability among patients with different etiologies was tested by one-way ANOVA. The area under the receiver operating characteristic curve (ROC) was calculated to compare the effectiveness between IPN and MODS prediction. A P value beneath 0.05 was considered statistically significant. All data were analyzed with commercial software (SPSS, Chicago, IL, version 19.0).
Results Characteristic of the Patients The general characteristic of 41 SAP patients are shown in Table 2. The etiology of pancreatitis was biliary in 20 patients (50 %), alcoholic in 7 patients (17 %), hyperlipemic in 9 patients (22 %) and due to other factors in 5 patients (12 %). Of 41 enrolled patients, 16 patients (39 %) developed IPN, 5 of them were treated operatively and 10 of them were treated with CT-guided percutaneous drainage only. A total of 9 patients developed MODS, and 4 patients died. The overall mortality was 9.8 %.
IPN infective pancreatic necrosis, APACHE acute physiology and chronic health evaluation, PCT procalcitonin, MODS multiple organ dysfunction syndrome, ICU intensive care unit
Heart Rate Variability in SAP Patients Presented with Different Clinical Variables As shown in Table 3, no difference was found in patients with different etiologies. In SAP patients who developed IPN or MODS, lower LF and LF/HF levels were found as compared to patients without these complications (Tables 4, 5). Moreover, nHF levels were significantly higher in patients who developed IPN or MODS (P \ 0.01). nLF levels were lower in MODS patients than patients who did not present with MODS. But nLF levels did not show any difference in patients with or without IPN. Heart rate variability levels of survivors and non-survivors are set out in Table 6. Values of nLF and LF/HF in non-survivals were lower, whereas nHF was significantly higher than those of survivals. No difference was found in time domain indexes. Comparison of Heart Rate Variability, APACHE II Score, and PCT in Predicting IPN and MODS Values of nHF and LF/HF for the prediction of IPN and MODS were analyzed by ROC curve. Figure 1 shows the ROC curves of nHF and LF/HF on admission in the prediction of IPN or MODS in SAP patients. Areas under ROC curves of nHF and LF/HF in prediction of IPN were 0.927 and 0.821, respectively. It seems that both nHF and LF/HF were superior to APACHE II score (AUC = 0.785) and PCT (AUC = 0.709) for the prediction of IPN. In the
123
Dig Dis Sci Table 3 Comparison of heart rate variability among SAP patients with different etiologies Gallstones SDNN(ms) RMSSD(ms) TP(ms2) 2
VLF(ms ) LF(ms2)
Alcohol
50.8 (19.4–110.2)
Hyperlipidemia
Others
P value
37.8 (21.3–72.6)
41.1 (17.9–76.4)
28 (13.4–97.9)
40.1 (10.4–84)
20.5 (12.7–88.9)
308.2 (127.9–920.5)
81.9 (66.7–865)
101.7 (28.2–463.7)
55.5 (12.5–511)
37.8 (8.1–302.4)
68.7 (51.6–107)
75.7 (13.6–164.4)
8.3 (2.7–144.6)
67.1 (1.5–334.5)
24 (19.6–63.4)
0.376
3.8 (2.7–10)
0.302
182.8 (22.3–1,863.8)
79 (59.4–280.5)
0.116
61.6 (37–406.6)
0.161
159.1 (94.1–311.2)
0.435 0.67
HF(ms2)
11 (4.8–42.8)
13.6 (2.1–17.7)
11.1 (7–103.3)
nLF(nu)
36.8 (17.5–73.2)
24.3 (6.6–57.8)
21.2 (3.5–52.5)
45.9 (19.8–64.3)
nHF(nu)
14.7 (4.9–24.2)
25.5 (4.1–33.7)
6.2 (4.6–26.8)
4.2 (3.4–14.4)
0.625
2.9 (0.2–10)
3.8 (0.7–6.7)
6.1 (3.4–15.3)
0.237
LF/HF
3 (2.1–6.1)
Table 4 Comparison of heart rate variability between patients with or without infected pancreatic necrosis IPN (n = 16)
Non-IPN (n = 25)
0.652
Table 6 Heart rate variability in non-survival patients Non-Survival (n = 4)
P value
Survival (n = 37)
P value
SDNN (ms)
35.5 (22.9–115.6)
64 (22.9–93.1)
0.612
SDNN (ms)
26.2 (8.2–108.7)
52 (25.1–93.1)
0.312
RMSSD (ms)
30.1 (10.2–99.8)
57.1 (17.9–85.2)
0.810
RMSSD (ms)
21.8 (6.7–175.9)
45.3 (15.9–85.2)
0.272
219.4 (35.8–893.2)
205.6 (94.6–968.2)
0.297
TP (ms2)
104.1 (32.5–820)
259.9 (88.2–921)
0.292
88.3 (7.8–710.5)
67.6 (36.7–271.6)
0.894
VLF (ms2)
8.7 (1.7–594.6)
70.4 (30.2–409.2)
0.124
67.6 (21.8–156)
0.034
LF (ms2)
2.6 (0.9–73.6)
53.8 (9.1–138.6)
0.083
12.8 (5.3–34.7)
10.6 (3.5–47.2)
0.575
HF (ms2)
26.7 (12.8–128.7)
11 (3.5–40.9)
0.135
nLF (nu)
25.2 (4.6–86.3)
39.5 (18–58.6)
0.612
nLF (nu)
3.2 (2.2–28.7)
39.5 (17.6–64.9)
0.020
nHF (nu)
30.6 (16.5–48)
4.6 (3.8–9)
\0.001
nHF (nu)
45.4 (26.6–76.1)
6.2 (4.1–21.2)
0.006
0.1 (0.04–0.5)
4 (2.4–7.7)
0.002
TP (ms2) 2
VLF (ms ) LF (ms2) HF(ms2)
LF/HF
9.1 (1.37–109.6)
1.0 (0.2–3.6)
4.6 (2.7–9.5)
0.001
LF/HF
IPN infective pancreatic necrosis Table 5 Comparison of heart rate variability between patients with or without multiple organ dysfunction syndromes (MODS) MODS (n = 9)
Non-MODS (n = 32)
P value
SDNN (ms)
30.4 (24–96.4)
50.9 (22.1–95.1)
0.219
RMSSD (ms)
15.5 (10.5–126.6)
54.3 (17.7–85.8)
0.270
TP (ms2) VLF (ms2)
126.3 (22.1–1,014.2) 232.8 (91.6–836.5)
0.395
12.5 (3.3–590.6)
72.4 (37.8–409.8)
0.078
2.7 (0.7–82.2)
60.7 (19.6–141.6)
0.036
HF (ms ) nLF (nu)
17.7 (9.4–42.7) 4.5 (2.3–31.5)
10.8 (3.4–42.8) 42 (17.9–64.9)
0.291 0.020
nHF (nu)
31.1 (24.4–52.8)
5.6 (4.1–17)
0.002
0.6 (0.06–1.2)
4.3 (2.7–8.5)
\0.001
LF (ms2) 2
LF/HF
Cut off Point, Sensitivity, Specificity, PPV and NPV of Heart Rate Variability, APACHE II Score and PCT for Early Assessment of IPN and MODS As shown in Table 8, in the prediction of IPN, the optimum cut off values of nHF, LF/HF, APACHE II score and PCT were 14 n.u, 1.2, 13.5 and 1.7 ng/ml, respectively. The optimum cut off values of nHF, LF/HF, APACHE II score and PCT for prediction of MODS were 21.2 n.u, 1.7, 15.5 and 1.7 ng/ml (Table 9). LF/HF showed the highest sensitivity of 89 % and highest specificity of 94 % in the prediction of MODS.
Discussion
MODS multiple organ dysfunction syndrome
prediction of MODS, areas under ROC curves of nHF and LF/HF were 0.837 and 0.906 (Fig. 1b). Effectiveness of nHF and LF/HF on differentiating of MODS from nonMODS was comparable with APACHE II score (AUC = 0.899), and better than PCT (AUC = 0.722) (Table 7).
123
In this study, we demonstrated that short-term spectral heart rate variability analysis performed within 24 h after admissions could identify SAP patients who were at risks of developing IPN, MODS and in-hospital mortality. We also found that LF and LF/HF levels were decreased while nHF levels were increased in patients who presented with IPN. In patients who developed MODS, lower LF, nLF and LF/HF levels and significantly higher nHF levels were
a
b
1.0
1.0
0.8
0.8
0.6
0.6
Sensitivity
Fig. 1 The ROC curves of nHF and LF/HF in predicting the development of infective pancreatic necrosis and MODS in SAP patients. a ROC curve of nHF and LF/HF in the prediction infective pancreatic necrosis. b ROC curve of nHF and LF/HF in the prediction of MODS
Sensitivity
Dig Dis Sci
0.4
0.2
0.4
0.2
nHF
nHF
LF/HF
LF/HF
0.0
0.0 0.0
0.2
0.4
0.6
0.8
1.0
0.0
0.2
1-Specificity
Table 7 Area under receiver operating characteristic curve (AUROC) for heart rate variability, PCT and APACHE II score in the prediction of infective pancreatic necrosis and MODS Variables
AUROC for IPN
0.4
0.6
0.8
1.0
1-Specificity
Table 9 Values of heart rate variability, APACHE II and PCT in the prediction of MODS
AUROC for MODS
Cut off value
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
nHF
0.927
0.837
nHF
C21.2n.u
89
84
62
96
LF/HF
0.821
0.906
LF/HF
B1.7
89
94
80
97
PCT
0.709
0.722
PCT
C1.7 ng/mL
56
88
56
88
APACHE II
0.785
0.899
APACHE II
C15.5
89
78
53
96
IPN infective pancreatic necrosis, PCT procalcitonin, APACHE acute physiology and chronic health evaluation, MODS multiple organ dysfunction syndrome Table 8 Values of heart rate variability, APACHE II and PCT in the prediction of infected pancreatic necrosis Cut off value
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
nHF
C14.0n.u
94
84
79
LF/HF
B1.2
89
94
100
78
PCT
C1.7 ng/mL
44
92
67
90
APACHE II
C13.5
88
72
78
72
95
PPV positive predictive value, NPV negative predictive value, PCT procalcitonin, APACHE acute physiology and chronic health evaluation
found as compared to non-MODS patients. It seems that the autonomic balance tilted toward sympathetic suppression in SAP patients who further developed pancreatic infection or MODS. Moreover, nHF could be a great predictor of pancreatic infection and MODS, with the receiver operating characteristic area of 0.927 and 0.837, respectively. The relationship between heart rate variability and clinical outcomes was also analyzed. Non-survivors patients had lower levels of nLF and LF/HF and significantly higher levels of nHF. These results suggested that SAP patients who suffered a more severe course have
MODS multiple organ dysfunction syndrome, PPV positive predictive value, NPV negative predictive value, PCT procalcitonin, APACHE II acute physiology and chronic health evaluation
sympathetic suppression in the early phase and heart rate variability proved to be a good predictor of SAP complications. Systemic inflammation is a normal response to disturbanced homeostasis and characterized by the endocrine release of different cytokines, such as TNF-a, IL-1, IL-6 and many others. These cytokines may induce activation of brain-derived neuroendocrine-immune interactions [19]. Neuroendocrine pathways, such as the sympathetic nervous system or parasympathetic nervous system, in turn, have some impact on the immune systems, through a-2 adrenoreceptor stimulation or the cholinergic anti-inflammatory pathway [20]. In the healthy state, there is some degree of stochastic variability in physiological variables, such as heart rate (heart rate variability). This variability is a measure of complexity that accompanies healthy systems and has been suggested to be responsible for their great adaptability and functionality related to pathologic systems [21]. Studying physiological signals of patients can offer us some ‘‘hidden’’ information of the progressions of diseases. Heart rate variability is a widely used approach to measure autonomic activity in humans. Heart rate is controlled by action potentials transmitted via the autonomic nerve to the sinoatrial node of the heart, where vagus nerve-dependent
123
Dig Dis Sci
acetylcholine release essentially ‘‘prolongs’’ the time to the next heart beat, thus slowing the pulse [20]. Heart rate variability is the time differences between successive heart beats, reflecting the balance of the autonomic nervous system regulation of heart rate. A variety of research has established two clear frequency bands in heart rate signals. These two bands include high frequency oscillations, between 0.15 and 0.4 Hz that are associated with vagus nerve activity, and bands with a lower frequency range, below 0.15 Hz, correlated with sympathetically and vagally mediated control of heart rate [19]. Heart rate variability has been strongly correlated with morbidity and mortality from diverse diseases [22]. Research has shown a link between heart rate variability and inflammatory markers in sepsis patients [23]. In fact, the loss of heart rate variability reflects ‘‘decomplexification’’ of the human physiology because connections among cells, organs and tissues are uncoupled, and this is proportional to the severity of diseases, especially critical illnesses; the recovery of heart rate variability can be transiently associated with clinical improvement and survival [24]. The first week of SAP patients always accompanies systemic inflammatory response syndrome. The cytokines released during this period could induce remote organ failure [25]. Concerning the studies mentioned above, autonomic nerve activity may play a role in the inflammatory response and disease severity in SAP patients. Heart rate variability might identify some ‘‘hidden’’ information in the early phase of SAP. SAP is a protean disease involving patients with various clinical courses and mortalities. IPN and MODS are two determinants of mortality [2]. Therefore, accurate prediction of IPN and MODS in the initial phase of SAP is valuable to manage appropriable treatments. A lot of work has been done to use heart rate variability for early diagnosis of infection in infants. Griffin, Moorman and colleagues developed heart rate characteristics to assess heart rate variability in infants [26, 27]. They found heart rate characteristics correlated well with sepsis, and reduced heart rate variability could predict sepsis before clinical diagnosis. Andrea Bravi et al. [28] studied 17 adult bone transplant patients, their results showed that heart rate variability was capable of tracking the development of sepsis, and the VLF was used to classify the probability of developing sepsis. Heart rate variability -based early diagnosis of infection was also investigated in post-strokes patients. Gunther A demonstrated that heart rate variability obtained within 48 h of hospital admission, a decreased nLF and LF/HF ratio and increased nHF could serve as early markers of infection in post-stroke patients [29]. Our results corroborated with the studies mentioned above. We found that LF and LF/HF ratio levels were decreased while nHF levels were increased in patients who developed
123
pancreatic infection as compared to non-infection patients. In addition, nHF and LF/HF could be used as early predictors of pancreatic infection in SAP patients. MODS is the sequential failure of at least three organs following a trigger event. Godin and Buchman proposed those neurally mediated organ interactions was ‘‘uncoupling’’ during MODS [30]. The toxins and mediators released in MODS might influence the cardiac cellular signal transduction, nerve connection and further leading to autonomic balance impairment. Studies by Hendrik Schmidt suggested heart rate variability was attenuated in MODS patients, and InVLF could predict 28-day mortality [31]. Our results also confirmed that autonomic balance titled toward sympathetic suppression in SAP patients who further presented MODS, and this phenomenon could be a good marker of MODS. Because timely and accurate evaluation of SAP severity is of major importance, several scoring systems and biomarkers have been studied for severity assessment. The most commonly used scoring system is APACHE II [32]. Our results showed that the frequency domain indexes— nHF and LF/HF could predict IPN and MODS comparable to APACHE II. Recently, increasing studies focused on the predictive value of PCT in SAP patients have shown that PCT is a reliable marker of pancreatic infection [33]. The ROC analysis of our study suggested that heart rate variability was superior to PCT in the prediction of both pancreatic infection and MODS. In addition, the measurement of heart rate variability is a bedside, noninvasive, low-cost and easy-to-perform method [24]. It seems that heart rate variability could probably replace PCT as a prognostic marker in SAP patients. To date, there are studies describing that a reduction of heart rate variability is correlated with the severity in sepsis patients, in critical ill patients and in MODS patients [9, 30, 31]. However, we did not encounter any published report focusing on the relationship between heart rate variability and SAP. Our study is one of the first to explore the role of heart rate variability in SAP patients. However, it should be noted that our study has several limitations. First, heart rate variability analysis cannot be performed in patients who have cardiac arrhythmia or using cardiac pacemaker. This may limit the applicability of heart rate variability to broader spectrum in SAP patients. But cardiac arrhythmia and cardiac pacing are not common in SAP patients, and none of the patients were excluded from our study for this reason. Second, we did not find any relationship between time domain indexes and illness severity, probably because the short-term recording of ECG. Time domain analysis usually requires 24 h of ECG recording, which is too long for ICU patients. Besides, vasopressor, sedative drugs used before or during data recording may substantially affect autonomic activity. It is unethical to interrupt these drugs
Dig Dis Sci
used in clinics for the sake of this study. However, studies have suggested that use of sedation or catecholamine do not affect sympatho-vagal balance in patients [34]. Despite the limitations mentioned above, our study does suggest heart rate variability obtained at admission could well predict disease severity in SAP patients. Conclusively, our study demonstrated that SAP patients who were at risks of pancreatic infection and MODS were characterized by sympathetic suppression in acute phase. We also confirmed that heart rate variability might be of great prognostic importance in SAP. Acknowledgments This study was supported by a grant from The National Natural Science Foundation of China, No. 81170438. Conflict of interest
None.
References 1. Al Mofleh IA. Severe acute pancreatitis: pathogenetic aspects and prognostic factors. World J Gastroenterol. 2008;14:675–684. 2. Petrov MS, Shanbhag S, Chakraborty M, Phillips ARJ, Windsor JA. Organ failure and infection of pancreatic necrosis as determinants of mortality in patients with acute pancreatitis. Gastroenterology. 2010;139:813–820. 3. Vege SS, Gardner TB, Chari ST, et al. Low mortality and high morbidity in severe acute pancreatitis without organ failure: a case for revising the Atlanta classification to include ‘‘moderately severe acute pancreatitis’’. Am J Gastroenterol. 2009;104: 710–715. 4. Aydemir M, Yazisiz V, Basarici I et al. Cardiac autonomic profile in rheumatoid arthritis and systemic lupus erythematosus. Lupus. 2010;19:255–261. 5. Garrard CS, Kontoyannis DA, Piepoli M. Spectral analysis of heart rate variability in the sepsis syndrome. Clin Auton Res. 1993;3:5–13. 6. Sharma P, Makharia GK, Ahuja V, Dwivedi SN, Deepak KK. Autonomic dysfunctions in patients with inflammatory bowel disease in clinical remission. Dig Dis Sci. 2009;54:853–861. 7. Chen WL, Chen JH, Huang CC, Kuo CD, Huang CI, Lee LS. Heart rate variability measures as predictors of in-hospital mortality in ED patients with sepsis. Am J Emerg Med. 2008;26:395–401. 8. Ahmad S, Tejuja A, Newman KD, Zarychanski R, Seely AJ. Clinical review: a review and analysis of heart rate variability and the diagnosis and prognosis of infection. Crit Care. 2009;13:232. 9. Pontet J, Contreras P, Curbelo A, et al. Heart rate variability as early marker of multiple organ dysfunction syndrome in septic patients. J Crit Care. 2003;18:156–163. 10. Bradley EL 3rd. A clinically based classification system for acute pancreatitis. Summary of the International Symposium on Acute Pancreatitis, Atlanta, Ga, September 11 through 13, 1992. Arch Surg. 1993;128:586–590. 11. Heinrich S, Schafer M, Rousson V, Clavien PA. Evidence-based treatment of acute pancreatitis: a look at established paradigms. Ann Surg. 2006;243:154–168. 12. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13:818–829. 13. Moreno R, Vincent JL, Matos R, et al. The use of maximum SOFA score to quantify organ dysfunction/failure in intensive
14.
15.
16.
17.
18.
19.
20.
21. 22. 23.
24.
25. 26.
27.
28.
29.
30.
31.
32.
care. Results of a prospective, multicentre study. Working Group on Sepsis related Problems of the ESICM. Intensive Care Med. 1999;25:686–696. Dellinger EP, Forsmark CE, Layer P, et al. Determinant-based classification of acute pancreatitis severity: an international multidisciplinary consultation. Ann Surg. 2012;256:875–880. Vincent JL, de Mendonca A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on ‘‘sepsis-related problems’’ of the European Society of Intensive Care Medicine. Crit Care Med. 1998;26:1793–1800. North American Society of Pacing and Electrophysiology, Task Force of the European Society of cardiology. Heart rate variability: standards of measurement, physiological interpretation and clinical use. Circulation. 1996;93:1043–1065. Kleiger RE, Stein PK, Bigger JT Jr. Heart rate variability: measurement and clinical utility. Ann Noninvasive Electrocardiol.. 2005;10:88–101. Pomeranz B, Macaulay RJ, Caudill MA, et al. Assessment of autonomic function in humans by heart rate spectral analysis. Am J Physiol. 1985;248:H151–H153. Papaioannou V, Pneumatikos I, Maglaveras N. Association of heart rate variability and inflammatory response in patients with cardiovascular disease: current strengths and limitations. Front Physiol. 2013;4:174. Huston JM, Tracey KJ. The pulse of inflammation: heart rate variability, the cholinergic anti-inflammatory pathway and implications for therapy. J Intern Med. 2011;269:45–53. Buchman TG. The community of the self. Nature. 2002;420: 246–251. Schmidt HB, Werdan K, Mu¨ller-Werdan U. Autonomic dysfunction in ICU patient. Curr Opin Crit Care.. 2001;7:314–322. Papaioannou VE, Dragoumanis C, Theodorou V, Gargaretas C, Pneumatikos I. Relation of heart rate variability to serum levels of C-reactive protein, interleukin 6, and 10 in patients with sepsis and septic shock. J Crit Care. 2009;24:6251–6257. Mazzeo AT, La Monaca E, Di Leo R, Vita G, Santamaria LB. Heart rate variability: a diagnostic and prognostic tool in anesthesia and intensive care. Acta Anaesthesiol Scand. 2011;55: 797–811. Pandol SJ, Saluja AK, Imrie CW, Banks PA. Acute pancreatitis: bench to the bedside. Gastroenterology. 2007;132:1127–1151. Griffin MP, Moorman JR. Toward the early diagnosis of neonatal sepsis and sepsis-like illness using novel heart rate analysis. Pediatrics. 2001;107:97–104. Griffin MP, O’Shea TM, Bissonette EA, Harrell FE Jr, Lake DE, Moorman JR. Abnormal heart rate characteristics preceding neonatal sepsis and sepsis-like illness. Pediatr Res. 2003;53: 920–926. Bravi A, Green G, Longtin A, Seely AJ. Monitoring and identification of sepsis development through a composite measure of heart rate variability. PLoS ONE. 2012;7:e45666. Gunther A, Salzmann I, Nowack S, et al. Heart rate variability—a potential early marker of sub-acute post-stroke infections. Acta Neurol Scand. 2012;126:189–196. Godin PJ, Buchman TG. Uncoupling of biological oscillators: a complementary hypothesis concerning the pathogenesis of multiple organ dysfunction syndrome. Crit Care Med. 1996;24: 1107–1116. Schmidt H, Hoyer D, Hennen R, et al. Autonomic dysfunction predicts both 1- and 2-month mortality in middle-aged patients with multiple organ dysfunction syndrome. Crit Care Med. 2008;36:967–970. Kaya E, Dervisoglu A, Polat C. Evaluation of diagnostic findings and scoring systems in outcome prediction in acute pancreatitis. World J Gastroenterol. 2007;13:3090–3094.
123
Dig Dis Sci 33. Rau BM, Kemppainen EA, Gumbs AA, et al. Early assessment of pancreatic infections and overall prognosis in severe acute pancreatitis by procalcitonin (PCT): a prospective international multicenter study. Ann Surg. 2007;245:745–754.
123
34. Schmidt H, Muller-Werdan U, Hoffmann T, et al. Autonomic dysfunction predicts mortality in patients with multiple organ dysfunction syndrome of different age groups. Crit Care Med. 2005;33:1994–2002.
ORIGINAL ARTICLE
Role of Heart Rate Variability in Predicting the Severity of Severe Acute Pancreatitis Luyao Zhang • Jing Zhou • Lu Ke • Yao Nie • Zhihui Tong • Weiqin Li • Jieshou Li
Received: 21 August 2013 / Accepted: 28 April 2014 Ó Springer Science+Business Media New York 2014
Abstract Background Infected pancreatic necrosis (IPN) and multiple organ dysfunction syndrome (MODS) are major complications of acute pancreatitis which determine disease severity and outcome. Aims The aim of this study is to investigate the value of admission heart rate variability as a marker of IPN or MODS in severe acute pancreatitis (SAP) patients. Methods Forty-one SAP patients within 72 h of symptoms onset were included in this prospective observational study. General demographics, laboratory data and the acute physiology and chronic health evaluation (APACHE) II scores were recorded at admission. 5-minute ECG signals were obtained at the same time for heart rate variability analyses to assess SAP severity.
L. Zhang J. Zhou L. Ke Y. Nie Z. Tong W. Li (&) J. Li Research Institute of General Surgery, Jinling Hospital, Medical School of Nanjing University, 305 East Zhongshan Road, Nanjing 210002, Jiangsu Province, China e-mail: [email protected] L. Zhang e-mail: [email protected] J. Zhou e-mail: [email protected] L. Ke e-mail: [email protected] Y. Nie e-mail: [email protected] Z. Tong e-mail: [email protected] J. Li e-mail: [email protected]
Results The baseline heart rate variability measurements, levels of low frequency/high frequency (LF/HF) were significantly lower whereas high frequency norm (nHF) levels were significantly higher in patients who present with IPN and MODS or died (P \ 0.01). Low frequency (LF) levels were lower in patients who present with IPN or MODS as compared to patients without these complications. Levels of low frequency norm (nLF) were lower in MODS and nonsurvival patients. nHF and LF/HF were good predictors of IPN and MODS, superior to procalcitonin. nHF and LF/HF were better than APACHE II in predicting IPN and LF/HF showed superiority over APACHE II in the prediction of MODS. Conclusions Admission heart rate variability is a good marker of IPN and MODS in SAP patients. Keywords Heart rate variability Infected pancreatic necrosis Multiple organ dysfunction syndrome Severe acute pancreatitis
Introduction Acute pancreatitis is the inflammation of pancreas and usually takes a self-limiting course. However, about 10–20 % acute pancreatitis patients progress to SAP, which is characterized by intensive inflammatory response in the early phase and IPN in the later phase [1]. SAP patients accompanying infection of pancreatic necrosis or organ failure are at high risk of mortality. The mortality increases to approximately 30 % when IPN is present and could reach as high as 46 % when MODS is present [2, 3]. Therefore, early prediction and diagnosis of IPN and organ failure are crucial in determinations of further treatment and prognosis.
123
Dig Dis Sci
Autonomic dysfunction occurs in many inflammatory diseases, including rheumatoid arthritis, inflammatory bowel disease, sepsis, and has been proved to be correlated with their morbidity and mortality [4–7]. Autonomic function could be well evaluated through three different methods: heart rate variability, baroreflex sensitivity and chemoreflex sensitivity. Heart rate variability is an indirect estimator of autonomic modulation of cardiovascular system and most commonly used in clinical research. Studies have suggested heart rate variability analysis providing early diagnosis and successful prognostication of infection in critically ill patients [8]. More specifically, decreased LF variation, a sign of impaired sympathetic activity, has been shown to closely associate with presence and severity of systemic infection [8]. Besides, heart rate variability could also predict prognosis in critical ill patients. Julio Pontet found that a reduction of heart rate variability could precede onset of MODS in septic shock patients [9]. In healthy people, the power of LF/HF ascends along with sympathetic activities. LF/HF provides an index of the dynamics of cardiac sympatho-vagal balance. Increase in parasympathetic tone over sympathetic tone (decrease in LF/HF) was identified as powerful predictor of mortality in cardiac risk patients undergoing major non-cardiac surgery and neurosurgical critically ill patients [8]. However, scarce attention has been paid to autonomic function in SAP patients. In the present study, we aimed to investigate the possible association between heart rate variability and disease severity in the acute phase of SAP. Furthermore, we also attempted to see whether heart rate variability can predict two determinant factors of severity in SAP: infection of pancreatic necrosis and organ failure.
Materials and Methods Patients A total of 41 adult patients with a diagnosis of SAP admitted at surgical intensive care unit of Jinling hospital within 72 h of abdominal pain onset from January 2012 to January 2013 were consecutively enrolled in this study. Diagnostic criteria of SAP were defined according to the Atlanta criteria [10]. Patients who presented one or more of the following conditions were excluded: (1) chronic pancreatitis, (2) pre-existing organ failure, (3) received surgical treatment before admission, (4) persistent arrhythmia and (5) use of cardiac pacemaker. The study protocol was approved by the ethics committee of Jinling hospital, and informed consent was obtained from the patient or his/her legal guardian. All patients received standard treatment for SAP [11].
123
Data Collection Demographic information and clinical relevant data of the patients were recorded. The APACHE II [12]. and SOFA (sequential organ failure assessment) scores [13] were calculated within 24 h on ICU admission. According to the new definition brought up by Dellinger et al. [14], IPN was defined when at least one of the following was present: (1) gas bubbles within (peri)pancreatic necrosis on computed tomography, (2) a positive culture of (peri)pancreatic necrosis obtained by image-guided fine-needle aspiration and (3) a positive culture of (peri)pancreatic necrosis obtained during the first drainage and/or necrosectomy. Organ failure was confirmed in three organ systems (cardiovascular, respiratory and renal) when a score of two or more according to the SOFA score presented [15]. MODS was defined as at least two of these three organ failures existed. In all patients, plasma procalcitonin (PCT) levels were measured on admission by chemoluminescent immunoassay (LUMITEST-PCT, BRAHMS Diagnostica AG, Hennigsdorf, Germany) in our central laboratory. Heart Rate Variability Analysis Measurement of heart rate variability was performed within 24 h of ICU admission. A 5-min ECG was obtained at bedside between 9:00 and 11:00 am using a 12-lead electrocardiogram machine (EDAN, Shenzhen, China). Heart rate variability was analyzed according to the standards developed by the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology [16]. Normal beat intervals were identified from the lead II ECG recordings by the Matlab software (MathWorks, Natick, MA). All artifacts and ectopic beats were removed in accordance with the guidelines outlined by the Taskforce of the European Society of Cardiology. The results were analyzed with both ‘‘time domain’’ and ‘‘frequency domain’’ methodologies [17, 18]. We calculated SDNN, RMSSD in the time domain. RMSSD primarily reflected vagus nerve-mediated changes in heart rate. SDNN was a marker of overall heart rate variability. The power spectrum of the beat-to-beat intervals was generated using fast Fourier transformation. The indexes of frequency domain calculated included: (a) VLF, in the frequency range less than 0.04 Hz and a dubious measure which did not have a specific physiological meaning. (b) LF, in the range of 0.04–0.15 Hz, correlated with sympathetically and vagally mediated control. (c) HF, in the region of 0.15–0.4 Hz, reflected vagally mediated control of heart period. (d) the LF/HF ratio, was used as the index of sympatho-vagal balance. HF and LF were also measured in their normalized forms, which represented the relative value of
Dig Dis Sci Table 1 Time domain and frequency domain measures of heart rate variability
Table 2 Characteristic of 41 SAP patients Age(yeas)
47 (34–59)
Variable
Units
Description
Gender(male/female)
22/19
VLF
ms2
Power in very low frequency range (B0.04 Hz)
Etiology, n (%)
LF nLF
ms
2
n.u. 2
HF
ms
nHF
n.u.
LF/HF SDNN RMSSP
ms ms
Power in low frequency range (0.04–0.15 Hz)
Gallstones
20 (50 %)
LF power in normalized units
Alcohol
7 (17 %)
LF/(Total Power–VLF) 9 100
Hyperlipidemia
9 (22 %)
Power in high frequency range (0.15–0.4 Hz)
Others
5 (12 %)
HF power in normalized units
APACHE II score
14 (11–18)
HF/(Total Power–VLF) 9 100 Ratio LF (ms2)/HF (ms2)
Ranson score
5 (4–7)
PCT level on admission(ng/mL)
0.3 (0.1–1.2)
Standard deviation of normal RR intervals
Balthazar index
6 (4–7)
Square root of the mean of the sum of the squares of differences between adjacent NN intervals
MODS
9 (22.0 %)
Sepsis
2 (4.9 %)
Pressure support
5 (12 %)
Surgical intervention IPN
15 (37 %) 16 (39 %)
Mortality
4 (9.8 %)
Length of ICU stay
9 (3–11)
Length of hospital stay
17 (8–18)
each power component in proportion to the total power minus the VLF component. The nHF reflected the modulation of vagus nerve discharge caused by respiration, nLF was the expression of baroreceptor-mediated regulation and due to the contribution of parasympathetic and mainly sympathetic discharge. The description of indexes we evaluated were set out in Table 1. Statistical Analysis Data were presented with its median and interquartile range (25–75th percentiles). Descriptive data were described as a absolute numbers or a percentage number. Mann–Whitney U test was used to compare continuous variables. The difference of heart rate variability among patients with different etiologies was tested by one-way ANOVA. The area under the receiver operating characteristic curve (ROC) was calculated to compare the effectiveness between IPN and MODS prediction. A P value beneath 0.05 was considered statistically significant. All data were analyzed with commercial software (SPSS, Chicago, IL, version 19.0).
Results Characteristic of the Patients The general characteristic of 41 SAP patients are shown in Table 2. The etiology of pancreatitis was biliary in 20 patients (50 %), alcoholic in 7 patients (17 %), hyperlipemic in 9 patients (22 %) and due to other factors in 5 patients (12 %). Of 41 enrolled patients, 16 patients (39 %) developed IPN, 5 of them were treated operatively and 10 of them were treated with CT-guided percutaneous drainage only. A total of 9 patients developed MODS, and 4 patients died. The overall mortality was 9.8 %.
IPN infective pancreatic necrosis, APACHE acute physiology and chronic health evaluation, PCT procalcitonin, MODS multiple organ dysfunction syndrome, ICU intensive care unit
Heart Rate Variability in SAP Patients Presented with Different Clinical Variables As shown in Table 3, no difference was found in patients with different etiologies. In SAP patients who developed IPN or MODS, lower LF and LF/HF levels were found as compared to patients without these complications (Tables 4, 5). Moreover, nHF levels were significantly higher in patients who developed IPN or MODS (P \ 0.01). nLF levels were lower in MODS patients than patients who did not present with MODS. But nLF levels did not show any difference in patients with or without IPN. Heart rate variability levels of survivors and non-survivors are set out in Table 6. Values of nLF and LF/HF in non-survivals were lower, whereas nHF was significantly higher than those of survivals. No difference was found in time domain indexes. Comparison of Heart Rate Variability, APACHE II Score, and PCT in Predicting IPN and MODS Values of nHF and LF/HF for the prediction of IPN and MODS were analyzed by ROC curve. Figure 1 shows the ROC curves of nHF and LF/HF on admission in the prediction of IPN or MODS in SAP patients. Areas under ROC curves of nHF and LF/HF in prediction of IPN were 0.927 and 0.821, respectively. It seems that both nHF and LF/HF were superior to APACHE II score (AUC = 0.785) and PCT (AUC = 0.709) for the prediction of IPN. In the
123
Dig Dis Sci Table 3 Comparison of heart rate variability among SAP patients with different etiologies Gallstones SDNN(ms) RMSSD(ms) TP(ms2) 2
VLF(ms ) LF(ms2)
Alcohol
50.8 (19.4–110.2)
Hyperlipidemia
Others
P value
37.8 (21.3–72.6)
41.1 (17.9–76.4)
28 (13.4–97.9)
40.1 (10.4–84)
20.5 (12.7–88.9)
308.2 (127.9–920.5)
81.9 (66.7–865)
101.7 (28.2–463.7)
55.5 (12.5–511)
37.8 (8.1–302.4)
68.7 (51.6–107)
75.7 (13.6–164.4)
8.3 (2.7–144.6)
67.1 (1.5–334.5)
24 (19.6–63.4)
0.376
3.8 (2.7–10)
0.302
182.8 (22.3–1,863.8)
79 (59.4–280.5)
0.116
61.6 (37–406.6)
0.161
159.1 (94.1–311.2)
0.435 0.67
HF(ms2)
11 (4.8–42.8)
13.6 (2.1–17.7)
11.1 (7–103.3)
nLF(nu)
36.8 (17.5–73.2)
24.3 (6.6–57.8)
21.2 (3.5–52.5)
45.9 (19.8–64.3)
nHF(nu)
14.7 (4.9–24.2)
25.5 (4.1–33.7)
6.2 (4.6–26.8)
4.2 (3.4–14.4)
0.625
2.9 (0.2–10)
3.8 (0.7–6.7)
6.1 (3.4–15.3)
0.237
LF/HF
3 (2.1–6.1)
Table 4 Comparison of heart rate variability between patients with or without infected pancreatic necrosis IPN (n = 16)
Non-IPN (n = 25)
0.652
Table 6 Heart rate variability in non-survival patients Non-Survival (n = 4)
P value
Survival (n = 37)
P value
SDNN (ms)
35.5 (22.9–115.6)
64 (22.9–93.1)
0.612
SDNN (ms)
26.2 (8.2–108.7)
52 (25.1–93.1)
0.312
RMSSD (ms)
30.1 (10.2–99.8)
57.1 (17.9–85.2)
0.810
RMSSD (ms)
21.8 (6.7–175.9)
45.3 (15.9–85.2)
0.272
219.4 (35.8–893.2)
205.6 (94.6–968.2)
0.297
TP (ms2)
104.1 (32.5–820)
259.9 (88.2–921)
0.292
88.3 (7.8–710.5)
67.6 (36.7–271.6)
0.894
VLF (ms2)
8.7 (1.7–594.6)
70.4 (30.2–409.2)
0.124
67.6 (21.8–156)
0.034
LF (ms2)
2.6 (0.9–73.6)
53.8 (9.1–138.6)
0.083
12.8 (5.3–34.7)
10.6 (3.5–47.2)
0.575
HF (ms2)
26.7 (12.8–128.7)
11 (3.5–40.9)
0.135
nLF (nu)
25.2 (4.6–86.3)
39.5 (18–58.6)
0.612
nLF (nu)
3.2 (2.2–28.7)
39.5 (17.6–64.9)
0.020
nHF (nu)
30.6 (16.5–48)
4.6 (3.8–9)
\0.001
nHF (nu)
45.4 (26.6–76.1)
6.2 (4.1–21.2)
0.006
0.1 (0.04–0.5)
4 (2.4–7.7)
0.002
TP (ms2) 2
VLF (ms ) LF (ms2) HF(ms2)
LF/HF
9.1 (1.37–109.6)
1.0 (0.2–3.6)
4.6 (2.7–9.5)
0.001
LF/HF
IPN infective pancreatic necrosis Table 5 Comparison of heart rate variability between patients with or without multiple organ dysfunction syndromes (MODS) MODS (n = 9)
Non-MODS (n = 32)
P value
SDNN (ms)
30.4 (24–96.4)
50.9 (22.1–95.1)
0.219
RMSSD (ms)
15.5 (10.5–126.6)
54.3 (17.7–85.8)
0.270
TP (ms2) VLF (ms2)
126.3 (22.1–1,014.2) 232.8 (91.6–836.5)
0.395
12.5 (3.3–590.6)
72.4 (37.8–409.8)
0.078
2.7 (0.7–82.2)
60.7 (19.6–141.6)
0.036
HF (ms ) nLF (nu)
17.7 (9.4–42.7) 4.5 (2.3–31.5)
10.8 (3.4–42.8) 42 (17.9–64.9)
0.291 0.020
nHF (nu)
31.1 (24.4–52.8)
5.6 (4.1–17)
0.002
0.6 (0.06–1.2)
4.3 (2.7–8.5)
\0.001
LF (ms2) 2
LF/HF
Cut off Point, Sensitivity, Specificity, PPV and NPV of Heart Rate Variability, APACHE II Score and PCT for Early Assessment of IPN and MODS As shown in Table 8, in the prediction of IPN, the optimum cut off values of nHF, LF/HF, APACHE II score and PCT were 14 n.u, 1.2, 13.5 and 1.7 ng/ml, respectively. The optimum cut off values of nHF, LF/HF, APACHE II score and PCT for prediction of MODS were 21.2 n.u, 1.7, 15.5 and 1.7 ng/ml (Table 9). LF/HF showed the highest sensitivity of 89 % and highest specificity of 94 % in the prediction of MODS.
Discussion
MODS multiple organ dysfunction syndrome
prediction of MODS, areas under ROC curves of nHF and LF/HF were 0.837 and 0.906 (Fig. 1b). Effectiveness of nHF and LF/HF on differentiating of MODS from nonMODS was comparable with APACHE II score (AUC = 0.899), and better than PCT (AUC = 0.722) (Table 7).
123
In this study, we demonstrated that short-term spectral heart rate variability analysis performed within 24 h after admissions could identify SAP patients who were at risks of developing IPN, MODS and in-hospital mortality. We also found that LF and LF/HF levels were decreased while nHF levels were increased in patients who presented with IPN. In patients who developed MODS, lower LF, nLF and LF/HF levels and significantly higher nHF levels were
a
b
1.0
1.0
0.8
0.8
0.6
0.6
Sensitivity
Fig. 1 The ROC curves of nHF and LF/HF in predicting the development of infective pancreatic necrosis and MODS in SAP patients. a ROC curve of nHF and LF/HF in the prediction infective pancreatic necrosis. b ROC curve of nHF and LF/HF in the prediction of MODS
Sensitivity
Dig Dis Sci
0.4
0.2
0.4
0.2
nHF
nHF
LF/HF
LF/HF
0.0
0.0 0.0
0.2
0.4
0.6
0.8
1.0
0.0
0.2
1-Specificity
Table 7 Area under receiver operating characteristic curve (AUROC) for heart rate variability, PCT and APACHE II score in the prediction of infective pancreatic necrosis and MODS Variables
AUROC for IPN
0.4
0.6
0.8
1.0
1-Specificity
Table 9 Values of heart rate variability, APACHE II and PCT in the prediction of MODS
AUROC for MODS
Cut off value
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
nHF
0.927
0.837
nHF
C21.2n.u
89
84
62
96
LF/HF
0.821
0.906
LF/HF
B1.7
89
94
80
97
PCT
0.709
0.722
PCT
C1.7 ng/mL
56
88
56
88
APACHE II
0.785
0.899
APACHE II
C15.5
89
78
53
96
IPN infective pancreatic necrosis, PCT procalcitonin, APACHE acute physiology and chronic health evaluation, MODS multiple organ dysfunction syndrome Table 8 Values of heart rate variability, APACHE II and PCT in the prediction of infected pancreatic necrosis Cut off value
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
nHF
C14.0n.u
94
84
79
LF/HF
B1.2
89
94
100
78
PCT
C1.7 ng/mL
44
92
67
90
APACHE II
C13.5
88
72
78
72
95
PPV positive predictive value, NPV negative predictive value, PCT procalcitonin, APACHE acute physiology and chronic health evaluation
found as compared to non-MODS patients. It seems that the autonomic balance tilted toward sympathetic suppression in SAP patients who further developed pancreatic infection or MODS. Moreover, nHF could be a great predictor of pancreatic infection and MODS, with the receiver operating characteristic area of 0.927 and 0.837, respectively. The relationship between heart rate variability and clinical outcomes was also analyzed. Non-survivors patients had lower levels of nLF and LF/HF and significantly higher levels of nHF. These results suggested that SAP patients who suffered a more severe course have
MODS multiple organ dysfunction syndrome, PPV positive predictive value, NPV negative predictive value, PCT procalcitonin, APACHE II acute physiology and chronic health evaluation
sympathetic suppression in the early phase and heart rate variability proved to be a good predictor of SAP complications. Systemic inflammation is a normal response to disturbanced homeostasis and characterized by the endocrine release of different cytokines, such as TNF-a, IL-1, IL-6 and many others. These cytokines may induce activation of brain-derived neuroendocrine-immune interactions [19]. Neuroendocrine pathways, such as the sympathetic nervous system or parasympathetic nervous system, in turn, have some impact on the immune systems, through a-2 adrenoreceptor stimulation or the cholinergic anti-inflammatory pathway [20]. In the healthy state, there is some degree of stochastic variability in physiological variables, such as heart rate (heart rate variability). This variability is a measure of complexity that accompanies healthy systems and has been suggested to be responsible for their great adaptability and functionality related to pathologic systems [21]. Studying physiological signals of patients can offer us some ‘‘hidden’’ information of the progressions of diseases. Heart rate variability is a widely used approach to measure autonomic activity in humans. Heart rate is controlled by action potentials transmitted via the autonomic nerve to the sinoatrial node of the heart, where vagus nerve-dependent
123
Dig Dis Sci
acetylcholine release essentially ‘‘prolongs’’ the time to the next heart beat, thus slowing the pulse [20]. Heart rate variability is the time differences between successive heart beats, reflecting the balance of the autonomic nervous system regulation of heart rate. A variety of research has established two clear frequency bands in heart rate signals. These two bands include high frequency oscillations, between 0.15 and 0.4 Hz that are associated with vagus nerve activity, and bands with a lower frequency range, below 0.15 Hz, correlated with sympathetically and vagally mediated control of heart rate [19]. Heart rate variability has been strongly correlated with morbidity and mortality from diverse diseases [22]. Research has shown a link between heart rate variability and inflammatory markers in sepsis patients [23]. In fact, the loss of heart rate variability reflects ‘‘decomplexification’’ of the human physiology because connections among cells, organs and tissues are uncoupled, and this is proportional to the severity of diseases, especially critical illnesses; the recovery of heart rate variability can be transiently associated with clinical improvement and survival [24]. The first week of SAP patients always accompanies systemic inflammatory response syndrome. The cytokines released during this period could induce remote organ failure [25]. Concerning the studies mentioned above, autonomic nerve activity may play a role in the inflammatory response and disease severity in SAP patients. Heart rate variability might identify some ‘‘hidden’’ information in the early phase of SAP. SAP is a protean disease involving patients with various clinical courses and mortalities. IPN and MODS are two determinants of mortality [2]. Therefore, accurate prediction of IPN and MODS in the initial phase of SAP is valuable to manage appropriable treatments. A lot of work has been done to use heart rate variability for early diagnosis of infection in infants. Griffin, Moorman and colleagues developed heart rate characteristics to assess heart rate variability in infants [26, 27]. They found heart rate characteristics correlated well with sepsis, and reduced heart rate variability could predict sepsis before clinical diagnosis. Andrea Bravi et al. [28] studied 17 adult bone transplant patients, their results showed that heart rate variability was capable of tracking the development of sepsis, and the VLF was used to classify the probability of developing sepsis. Heart rate variability -based early diagnosis of infection was also investigated in post-strokes patients. Gunther A demonstrated that heart rate variability obtained within 48 h of hospital admission, a decreased nLF and LF/HF ratio and increased nHF could serve as early markers of infection in post-stroke patients [29]. Our results corroborated with the studies mentioned above. We found that LF and LF/HF ratio levels were decreased while nHF levels were increased in patients who developed
123
pancreatic infection as compared to non-infection patients. In addition, nHF and LF/HF could be used as early predictors of pancreatic infection in SAP patients. MODS is the sequential failure of at least three organs following a trigger event. Godin and Buchman proposed those neurally mediated organ interactions was ‘‘uncoupling’’ during MODS [30]. The toxins and mediators released in MODS might influence the cardiac cellular signal transduction, nerve connection and further leading to autonomic balance impairment. Studies by Hendrik Schmidt suggested heart rate variability was attenuated in MODS patients, and InVLF could predict 28-day mortality [31]. Our results also confirmed that autonomic balance titled toward sympathetic suppression in SAP patients who further presented MODS, and this phenomenon could be a good marker of MODS. Because timely and accurate evaluation of SAP severity is of major importance, several scoring systems and biomarkers have been studied for severity assessment. The most commonly used scoring system is APACHE II [32]. Our results showed that the frequency domain indexes— nHF and LF/HF could predict IPN and MODS comparable to APACHE II. Recently, increasing studies focused on the predictive value of PCT in SAP patients have shown that PCT is a reliable marker of pancreatic infection [33]. The ROC analysis of our study suggested that heart rate variability was superior to PCT in the prediction of both pancreatic infection and MODS. In addition, the measurement of heart rate variability is a bedside, noninvasive, low-cost and easy-to-perform method [24]. It seems that heart rate variability could probably replace PCT as a prognostic marker in SAP patients. To date, there are studies describing that a reduction of heart rate variability is correlated with the severity in sepsis patients, in critical ill patients and in MODS patients [9, 30, 31]. However, we did not encounter any published report focusing on the relationship between heart rate variability and SAP. Our study is one of the first to explore the role of heart rate variability in SAP patients. However, it should be noted that our study has several limitations. First, heart rate variability analysis cannot be performed in patients who have cardiac arrhythmia or using cardiac pacemaker. This may limit the applicability of heart rate variability to broader spectrum in SAP patients. But cardiac arrhythmia and cardiac pacing are not common in SAP patients, and none of the patients were excluded from our study for this reason. Second, we did not find any relationship between time domain indexes and illness severity, probably because the short-term recording of ECG. Time domain analysis usually requires 24 h of ECG recording, which is too long for ICU patients. Besides, vasopressor, sedative drugs used before or during data recording may substantially affect autonomic activity. It is unethical to interrupt these drugs
Dig Dis Sci
used in clinics for the sake of this study. However, studies have suggested that use of sedation or catecholamine do not affect sympatho-vagal balance in patients [34]. Despite the limitations mentioned above, our study does suggest heart rate variability obtained at admission could well predict disease severity in SAP patients. Conclusively, our study demonstrated that SAP patients who were at risks of pancreatic infection and MODS were characterized by sympathetic suppression in acute phase. We also confirmed that heart rate variability might be of great prognostic importance in SAP. Acknowledgments This study was supported by a grant from The National Natural Science Foundation of China, No. 81170438. Conflict of interest
None.
References 1. Al Mofleh IA. Severe acute pancreatitis: pathogenetic aspects and prognostic factors. World J Gastroenterol. 2008;14:675–684. 2. Petrov MS, Shanbhag S, Chakraborty M, Phillips ARJ, Windsor JA. Organ failure and infection of pancreatic necrosis as determinants of mortality in patients with acute pancreatitis. Gastroenterology. 2010;139:813–820. 3. Vege SS, Gardner TB, Chari ST, et al. Low mortality and high morbidity in severe acute pancreatitis without organ failure: a case for revising the Atlanta classification to include ‘‘moderately severe acute pancreatitis’’. Am J Gastroenterol. 2009;104: 710–715. 4. Aydemir M, Yazisiz V, Basarici I et al. Cardiac autonomic profile in rheumatoid arthritis and systemic lupus erythematosus. Lupus. 2010;19:255–261. 5. Garrard CS, Kontoyannis DA, Piepoli M. Spectral analysis of heart rate variability in the sepsis syndrome. Clin Auton Res. 1993;3:5–13. 6. Sharma P, Makharia GK, Ahuja V, Dwivedi SN, Deepak KK. Autonomic dysfunctions in patients with inflammatory bowel disease in clinical remission. Dig Dis Sci. 2009;54:853–861. 7. Chen WL, Chen JH, Huang CC, Kuo CD, Huang CI, Lee LS. Heart rate variability measures as predictors of in-hospital mortality in ED patients with sepsis. Am J Emerg Med. 2008;26:395–401. 8. Ahmad S, Tejuja A, Newman KD, Zarychanski R, Seely AJ. Clinical review: a review and analysis of heart rate variability and the diagnosis and prognosis of infection. Crit Care. 2009;13:232. 9. Pontet J, Contreras P, Curbelo A, et al. Heart rate variability as early marker of multiple organ dysfunction syndrome in septic patients. J Crit Care. 2003;18:156–163. 10. Bradley EL 3rd. A clinically based classification system for acute pancreatitis. Summary of the International Symposium on Acute Pancreatitis, Atlanta, Ga, September 11 through 13, 1992. Arch Surg. 1993;128:586–590. 11. Heinrich S, Schafer M, Rousson V, Clavien PA. Evidence-based treatment of acute pancreatitis: a look at established paradigms. Ann Surg. 2006;243:154–168. 12. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13:818–829. 13. Moreno R, Vincent JL, Matos R, et al. The use of maximum SOFA score to quantify organ dysfunction/failure in intensive
14.
15.
16.
17.
18.
19.
20.
21. 22. 23.
24.
25. 26.
27.
28.
29.
30.
31.
32.
care. Results of a prospective, multicentre study. Working Group on Sepsis related Problems of the ESICM. Intensive Care Med. 1999;25:686–696. Dellinger EP, Forsmark CE, Layer P, et al. Determinant-based classification of acute pancreatitis severity: an international multidisciplinary consultation. Ann Surg. 2012;256:875–880. Vincent JL, de Mendonca A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on ‘‘sepsis-related problems’’ of the European Society of Intensive Care Medicine. Crit Care Med. 1998;26:1793–1800. North American Society of Pacing and Electrophysiology, Task Force of the European Society of cardiology. Heart rate variability: standards of measurement, physiological interpretation and clinical use. Circulation. 1996;93:1043–1065. Kleiger RE, Stein PK, Bigger JT Jr. Heart rate variability: measurement and clinical utility. Ann Noninvasive Electrocardiol.. 2005;10:88–101. Pomeranz B, Macaulay RJ, Caudill MA, et al. Assessment of autonomic function in humans by heart rate spectral analysis. Am J Physiol. 1985;248:H151–H153. Papaioannou V, Pneumatikos I, Maglaveras N. Association of heart rate variability and inflammatory response in patients with cardiovascular disease: current strengths and limitations. Front Physiol. 2013;4:174. Huston JM, Tracey KJ. The pulse of inflammation: heart rate variability, the cholinergic anti-inflammatory pathway and implications for therapy. J Intern Med. 2011;269:45–53. Buchman TG. The community of the self. Nature. 2002;420: 246–251. Schmidt HB, Werdan K, Mu¨ller-Werdan U. Autonomic dysfunction in ICU patient. Curr Opin Crit Care.. 2001;7:314–322. Papaioannou VE, Dragoumanis C, Theodorou V, Gargaretas C, Pneumatikos I. Relation of heart rate variability to serum levels of C-reactive protein, interleukin 6, and 10 in patients with sepsis and septic shock. J Crit Care. 2009;24:6251–6257. Mazzeo AT, La Monaca E, Di Leo R, Vita G, Santamaria LB. Heart rate variability: a diagnostic and prognostic tool in anesthesia and intensive care. Acta Anaesthesiol Scand. 2011;55: 797–811. Pandol SJ, Saluja AK, Imrie CW, Banks PA. Acute pancreatitis: bench to the bedside. Gastroenterology. 2007;132:1127–1151. Griffin MP, Moorman JR. Toward the early diagnosis of neonatal sepsis and sepsis-like illness using novel heart rate analysis. Pediatrics. 2001;107:97–104. Griffin MP, O’Shea TM, Bissonette EA, Harrell FE Jr, Lake DE, Moorman JR. Abnormal heart rate characteristics preceding neonatal sepsis and sepsis-like illness. Pediatr Res. 2003;53: 920–926. Bravi A, Green G, Longtin A, Seely AJ. Monitoring and identification of sepsis development through a composite measure of heart rate variability. PLoS ONE. 2012;7:e45666. Gunther A, Salzmann I, Nowack S, et al. Heart rate variability—a potential early marker of sub-acute post-stroke infections. Acta Neurol Scand. 2012;126:189–196. Godin PJ, Buchman TG. Uncoupling of biological oscillators: a complementary hypothesis concerning the pathogenesis of multiple organ dysfunction syndrome. Crit Care Med. 1996;24: 1107–1116. Schmidt H, Hoyer D, Hennen R, et al. Autonomic dysfunction predicts both 1- and 2-month mortality in middle-aged patients with multiple organ dysfunction syndrome. Crit Care Med. 2008;36:967–970. Kaya E, Dervisoglu A, Polat C. Evaluation of diagnostic findings and scoring systems in outcome prediction in acute pancreatitis. World J Gastroenterol. 2007;13:3090–3094.
123
Dig Dis Sci 33. Rau BM, Kemppainen EA, Gumbs AA, et al. Early assessment of pancreatic infections and overall prognosis in severe acute pancreatitis by procalcitonin (PCT): a prospective international multicenter study. Ann Surg. 2007;245:745–754.
123
34. Schmidt H, Muller-Werdan U, Hoffmann T, et al. Autonomic dysfunction predicts mortality in patients with multiple organ dysfunction syndrome of different age groups. Crit Care Med. 2005;33:1994–2002.
