CARDIAC ARRHYTHMIAS AND SILENT MYOCARDIAL ISCHEMIA DURING HEMODIALYSIS

Abstract

Cardiac arrhythmias are noted in a significant proportion of chronic renal failure (CRF) patients on hemodialysis (HD), and may contribute to cardiovascular mortality. A number of factors have been implicated in the genesis of these arrhythmias. The role of silent myocardial ischemia (SMI), however, has not been evaluated systematically. We prospectively studied 38 unselected CRF patients on regular HD by continuous Holter monitoring starting 24 hours before HD, lasting through the dialysis session and continued for 20 hours thereafter. The recordings were analyzed for frequency, timing and severity of supraventricular and ventricular arrhythmias and SMI as identified by ST-segment depression. Ventricular arrhythmias during HD were noted in 11 (29%) patients (group I), and were potentially life-threatening (Lown Class III and IVa) in 13%. The remaining 27 patients (group II) had no ventricular arrhythmias during HD. There was no difference in the age, sex ratio, duration of HD, blood pressure, fluctuations in weight, hematocrit, predialysis creatinine, sodium, potassium, calcium or inorganic phosphate levels between patients in the two groups. The number of patients with clinical ischemic heart disease was significantly greater in group I. SMI was noted in 72% and 33% of group I and II patients respectively (p =0.026). 46% of those with and 25% of those without ST changes during HD developed ventricular arrthythmias during HD. Both SMI and ventricular arrhythmias were noted most frequently during the last hour of dialysis. Hypertension, diabetes mellitus and ischemic heart disease were observed more frequently amongst patients with SMI. Ventricular arrhythmias are detected in a significant proportion of CRF patients on HD. These are probably related to coronary artery disease since silent myocardial ischemia is also noted more frequently during HD in these patients. Further studies incorporating coronary angiography are needed in a larger number of patients to establish a definite causal relationship.

• Chronic renal failure
• Hemodialysis
• Cardiac arrhythmias
• Silent myocardial ischemia

INTRODUCTION

Renal failure and hemodialysis (HD) have a profound effect on the cardiovascular system. Cardiac causes account for 40–50% of all deaths in dialysis patients in the Western countries [1-2]. About 15–20% of these deaths occur suddenly and unexpectedly [1-5]. According to Sforzini et al. [[6]], one quarter of these deaths occur within one hour of onset of symptoms. The most important cause of this mortality is coronary artery disease leading to myocardial infarction and congestive cardiac failure [[7]]. Cardiac arrhythmias are noted in a significant proportion of dialysis patients [[3]], [8-10] and may contribute to the increased mortality. However, there is no agreement on whether or not the HD process itself is arrhythmogenic [7-10]. It would, therefore, be important to determine whether silent arrhythmias are indeed seen more frequently during HD and to identify those at increased risk for development of life threatening arrhythmias.

The factors that have been implicated in the genesis of these arrhythmias include abnormal plasma catecholamine levels [[11]], uremic cardiomyopathy [[12]], digitalis use [13-14] dialysate buffer [[10]] and secondary hyperparathyroidism [[15]]. In addition, rapid changes in the metabolic milieu that occur during the HD process itself may increase the risk of arrhythmias [[13]], [16-17]. The effect of myocardial ischemia during HD on the development of arrhythmias has not been examined. Alterations in the T wave and ST segment morphology are indicative of myocardial ischemia [18-20]. Ambulatory (Holter) electrocardiography is especially useful to record the frequency and duration of transient cardiac arrhythmias as well as myocardial ischemia during HD.

The present study was carried out to determine the incidence and type of cardiac arrhythmias in our chronic HD patients using 48-hour Holter monitoring starting a day prior to dialysis and to correlate these with the occurrence of painless myocardial ischemia as indicated by the ST segment abnormalities.

MATERIAL AND METHODS

A total of 38 unselected patients of chronic renal failure on HD were included in this prospective study with their informed consent. All patients were more than 18 years of age and had been clinically stable for at least two months before the study. None of the patients were on anti-arrhythmic drugs. The commonest cause of renal failure in these patients was chronic glomerulonephritis (28 cases), followed by diabetic nephropathy [[4]], chronic interstitial nephritis [[3]], autosomal dominant polycystic kidney disease [[2]] and obstructive uropathy [[1]].

Patients underwent detailed physical examination, standard 12-lead surface electrocardiogram, X-ray chest and M-mode and 2D-echocardiography. Body weight and blood pressure were recorded before and after HD. A predialysis venous blood sample was drawn and analysed for hematocrit, urea, creatinine, sodium, potassium, calcium and inorganic phosphate levels.

All patients received chronic HD treatment twice every week for a total of 8 hours/week, which is the common dialysis practice in India because of financial constraints. The HD was performed using cuprophane hollow fiber dialyzers (surface area 1–1.2 m²). Blood flow rate was maintained at 220–250 mL/min and the dialysate flow was 500 mL/min. The dialysis solution contained: sodium 136 mmol/L, chloride 102.75 mmol/L, potassium 2 mmol/L, calcium 3.25 mmol/L and acetate 37 mmol/L.

Holter Monitoring

Continuous Holter monitoring was started 24 hours before HD and continued through the dialysis for a further 24 hour period, giving a 20-hour post dialysis record. MV₁ (modified V₁) and MV₅ (modified V₅) leads were recorded on magnetic tapes and later analysed by Diagnostic Medical Instruments (DMI-1529) computer-assisted analyser.

The patients were not restricted in their activities during the period of recording. The arrhythmias that were noted included premature supraventricular beats (PSB), premature ventricular contractions (PVC) and complex ventricular arrhythmias (couplets or ventricular tachycardia). ST segment depression from the baseline was measured 80 msec after the J-point. Patients were instructed to keep diaries during the monitoring period to note the time of onset and severity of chest discomfort and nature of their activity at that time.

Definitions

Ventricular arrhythmias were classified according to the Lown and Grabosys [[21]] classification: Class 0=no PVC, Class I: <30 PVC/hour, Class II: >30 PVC/hour, Class III: polymorphic PVC, Class IVa: PVC with couplets, Class IVb: ventricular tachycardia (>3 PVC) and Class V: early PVC (R-R/QT <0.85). LVH was defined on electrocardiogram when the combined heights of S wave in V1 and R wave in V5 leads exceeded 3.5 mv and on echocardiography by demonstration of interventricular septal (IVS) or left ventricular posterior wall (LVPW) thickness > 12 mm. Left ventricular dysfunction was denoted by left ventricular ejection fraction <45% on echocardiography. Clinical ischemic heart disease was defined as a history of acute myocardial infarction or angina.

The Holter recording was considered positive for silent myocardial ischemia if the ST segment showed a >0.1 mV depression 80 msec after the J point that persisted for more than 60 milliseconds.

The Holter recordings were divided into the following four-hour phases for purposes of comparison:

• Period 1 : 20–24 hours before HD

• Period 2 : 4–8 hours before HD

• Period 3 : 4 hours before HD

• Period 4 : During HD

• Period 5 : 4 hours immediately after HD

• Period 6 : 8–12 hours after HD

Statistical Analysis

All parameters were expressed as mean ±SD. Differences in parameters were analysed using unpaired Student's ‘t’ test and Pearson's Chi-square test p value of <0.05 was considered significant.

RESULTS

Of the 38 patients studied, arrhythmias were noted during HD sessions in 11 (29%) cases (group 1). Complex ventricular arrhythmias of Lown's class III and IVa were seen in 5 (13.2%) of these patients, but none had class IVb or V arrhythmias. The remaining 27 patients were categorised as group 2. Occasional ventricular ectopic activity with a frequency of ≤3 beats/hour before or after HD was noted in 5 of these 27 cases, and this did not worsen during dialysis. Thus, a total of 16 (28.9%) cases showed ventricular arrhythmias either during HD or in the interdialytic period. The demographic characteristics of patients in both the groups are shown in Table 1. Four patients in group I and two in group II had ischemic heart disease (p=0.026) Of these, one in each group had ECG evidence of old inferior wall myocardial infarction, two in group I had old anteroseptal sub-endocardial myocardial infarction and one additional patient in this group had effort angina. One patient in each group was on digoxin for congestive cardiac failure.

Table 1. Demographic Characteristics of Patients with and Without Ventricular Arrhythmias During HD

	Group I	Group II	p value
Age (years)	45.6±14	39.6±12.3	>0.05
Sex Ratio M:F	10:1	21:6	>0.05
Duration of HD (weeks)	4.77±3.92	3.74±1.02	>0.05
Blood pressure (mm of Hg)
Systolic B	150±26.8	146.3±19	>0.05
A	142.7±20	140±15.2	>0.05
Diastolic B	92.4±10.8	91.1±10.1	>0.05
A	87.3±11	85.1±8	>0.05
Weight gain (kg)	1.8±1.5	1.63±1	>0.05
Weight loss (kg)	1.88±1.6	1.34±0.8	>0.05

B: before HD, A: after HD

Patients with and without ventricular arrhythmias had comparable degrees of hypertension and fluctuations of weight before and after dialysis (Table 1). The pre-dialysis blood hematocrit, sodium, potassium, calcium, phosphorus, urea and creatinine did not differ significantly in the two groups (Table 2). Patients in group 1 had a higher proportion of patients with LVH and left ventricular dysfunction, but the difference failed to reach statistical significance (Table 3).

Table 2. Predialysis Laboratory Data of Patients with and Without Ventricular Arrhythmias During Dialysis

Parameter	Group I	Group II	p value
Hematocrit(%)	22±5.3	21.2±3.4	>0.05
Blood urea (mmol/L)	43.9±9.9	52.86±15.96	>0.05
Serum creatinine (μmol/L)	836±184.6	959.2±352	>0.05
Serum sodium (mmol/L)	128±7.1	128±6.3	>0.05
Serum potassium (mmol/L)	4.5±0.6	4.8±0.9	>0.05
Serum calcium (mmol/L)	1.9±0.37	1.81±0.27	>0.05
Serum phospate (nmol/L)	1.95±0.48	2.72±0.48	>0.05

Table 3. Echocardiographic Parameters of Patients with and Without Ventricular Arrhythmias

Parameter	Group I	Group II	p value
Left ventricular free wall thickness (cm)	1.6±0.5	1.4±0.3	>0.05
Left ventricular ejection Fraction (%)	45.9±12.6	56.1±13.7	>0.05
Left ventricular hypertrophy	8 (72.7%)	17 (62.9%)	>0.05
Left ventricular dysfunction	5 (45.5%)	9 (33.3%)	>0.05

Symptomatic hypotension during HD was seen in two patients in group II but was not observed in any patient with ventricular arrhythmias. Two patients in group I had experienced angina pectoris during an earlier dialysis but had remained asymptomatic during the index dialysis.

An hourly analysis during the period of HD showed that the frequency of ventricular ectopics increased in the first hour of HD and was maximum during the 4^th hour of dialysis (Fig 1). There was no significant increase in the ectopic activity in the immediate or late post HD periods compared to corresponding pre dialysis periods (Fig 2).

Figure 1. Hourly frequency of VPCs during different 4 hour periods.

Figure 2. Hourly VPC rates during different study periods Group I patients.

Supraventricular arrhythmias were noted in 20 (52.6%) cases. These were noted during HD and in the interdialytic period in 11 cases and during interdialytic period only in nine cases. In all but two instances, these consisted of isolated paroxysmal supraventricular ectopics, with a median frequency of 2/hour. Two patients showed transient atrial fibrillation during the interdialytic period lasting from 3–34 minutes.

On follow up, two patients from group I died suddenly. Complex ventricular arrhythmias had been recorded in one of these during an earlier HD.

ST Segment Changes

A total of 165 episodes of dynamic ST segment changes were noted in 17 patients. Of these, seven patients exhibited these changes only during HD, 4 during the interdialytic period and six both during HD and the interdialytic period. Thus, a total of 13 patients exhibited ST changes during HD. The duration of these episodes ranged from 1 – 177 minutes with a mean of 4.3 minutes. The maximum ST segment depression was 4 mm. Eight patients experienced 1–5 episodes during the study period, three had 6–15 episodes and six were recorded to have > 16 episodes. ST segment abnormalities were most frequent in the last hour of HD (69%) and persisted up to 10 hours after HD in three patients. None of the patients complained of any symptom at the times when ST changes were recorded. An analysis of risk factors showed that while the age and sex ratio were the same in patients with and without ST segment changes; hypertension, diabetes mellitus and ischemic heart disease were more frequently observed in patients who showed ST changes (Table 4).

Table 4. Comparison of Different Parameters in Patients with and Without ST Changes

Parameter	ST changes present (n=17)	ST changes absent (n=21)	p value
Age	45±14	35±8	>0.05
Sex ratio M:F	14.3	17:4	>0.05
Hypertension	10 (58.8%)	5 (23.8%)	0.028
Diabetes	5 (29.4%)	0	0.008
Ischemic heart disease	5 (29.4%)	1 (4.8%	0.038
LVEF(%)	52±10	51±11	>0.05
IVS(s)cm	1.2±0.08	1.2±0.06	>0.05
IVS(d) cm	1.1±0.07	1.1±0.08	>0.05
LVH	2	2	>0.05
LV dysfunction	2	2	>0.05
Hypotension	0	3	>0.05

Correlation Between Ventricular Arrhythmias and ST Segment Changes

ST segment changes were noted in 8 out of 11 (72.7%) group 1 and 9 of 27 (33.3%) group 2 patients (p=0.026). Out of the 13 patients who exhibited ST changes during dialysis, 6 (46%) also developed ventricular arrhythmias during HD. On the other hand, ventricular arrhythmias were noted in 25% of patients in whom the ST changes were restricted to the interdialytic period (p =0.48).

DISCUSSION

Changes in the blood volume, electrolytes and osmolality and the interaction of the patient's blood with the dialysis membrane would be expected to provide sufficient stimuli to induce arrhythmias during a HD session. However, the arrhythmogenic effect of dialysis has not been uniformly documented (Table 1). The incidence of arrhythmias amongst dialysis patients varies from 0–95% [[3]], [7-10], [[13]], [15-18], [[22]], [[25]]. We documented supraventricular and ventricular arrhythmias in 53% and 29% respectively amongst patients on chronic maintenance HD, with 13% patients experiencing potentially life-threatening arrhythmias. This variation is due to differences in the patient population, inclusion criteria, definition of arrhythmias and the method and duration for which the patients were studied. There are significant differences in the epidemiological pattern of ESRD in the western countries and developing nations like India. The ESRD population of the western nations is older and more likely to have diabetic nephropathy and hypertensive nephrosclerosis, conditions that are associated with an increased likelihood of cardiovascular abnormalities [1-2]. The patients in the present study, on the other hand, were younger and most had ESRD due to chronic glomerulonephritis. This population is a representative sample of ESRD patients in our country. The age, sex ratio and the underlying renal disease leading to ESRD is similar to that in larger series of patients reported earlier [23-24]. The dialysis duration of 8 hours/week as in the present study would be considered inadequate by western standards. However, this dialysis specification and frequency is similar to that administered to an overwhelming majority of patients in our country. The duration for which the patients had been on dialysis in the present study was only 4.04 months, with only two patients being on dialysis for more than 6 months. This is far less than that reported in the other studies, where the mean duration varied from 12–89 months [[3]], [7-10], [[13]], [15-18], [[22]], [[25]].

Some of the series [[7]], [[10]], [[16]] have included only those patients who had shown rhythm abnormalities on baseline EKG or during an earlier dialysis. We studied all patients undergoing HD in our unit. The prevalence of ventricular arrhythmias was significantly higher during the 4 hour HD period than other 4-hour periods before and after HD and high grade arrhythmias were noted only during HD. Other workers have also reported high frequency of arrhythmias during the dialysis period that gradually diminishes in the period immediately following HD [[3]], [[8]], [[13]], [16-17], [[25]]. Occurrence of arrhythmias before dialysis has been shown to confer an increased risk of development of arrhythmias during dialysis [[3]].

Out of the two patients who died suddenly, one was documented to have ventricular arrhythmia on an earlier occasion. The role of ventricular arrhythmias in increasing the susceptibility to a fatal outcome in CRF patients has been suggested, but not proved. According to the 1996 USRDS Report [[1]], cardiac arrhythmias were responsible for 6% of all deaths in dialysis patients. In another 17%, the cause of death was listed as ‘cardiac arrest, cause unknown’. Chazan et al. [[4]] reported that 25 of their 31 in-center cardiac deaths were related to ventricular fibrillation. Sforzini et al.[[6]] followed 127 patients who had been documented to have a high prevalence of ventricular arrhythmias during and after dialysis for four years. On univariate analysis, a PVC rate of >2/hr seemed to increase the mortality risk by a factor of 1.76, but the difference failed to reach statistical insignificance on multivariate analysis.

The pathogenesis of arrhythmias in CRF patients on HD is unclear. A number of possible factors have been suggested by different workers. These include advanced age [[3]], ischemic heart disease [[7]], [[16]], [[26]], dialysis induced decrement in serum potassium [[13]], [16-17] longer duration of dialysis [[9]], high dialysate calcium [[8]], [[27]], digitalis therapy [13-14], use of acetate as a buffer [9-10], increased plasma nonepinephrine [[11]], uremic cardiomyopathy [[8]], [[12]], autonomic neuropathy [28-29], increased parathyroid hormone concentration [[15]], changes in blood volume and left ventricular hypertrophy [[7]], [[28]]. Of these, the rate of change in serum potassium concentration during dialysis has been thought to be important in the genesis of dialysis induced arrhythmias. Redaelli et al. [[30]] demonstrated that the arrhythmogenic effect of HD was less when the plasma-dialysate K⁺ gradient was kept constant throughout HD instead of a fixed low dialysate K⁺ concentration. Similarly, Weber et al.[[17]] found that the post-dialysis K⁺ concentration was lower in the group of patients who had arrhythmias. Some recent studies have shown the QT interval is prolonged in CRF patients. HD induces a further increase in dispersion of the QTc values [[31]] and these patients are more likely to develop complex VPCs [[32]].

In our study, there was no difference in the age, sex ratio or body weight changes during dialysis or interdialytic period between the two groups. Similarly, the predialysis hematocrit, blood urea, creatinine, sodium, potassium, calcium and phosphate levels were similar in the two groups. We did not determine the postdialysis concentration of these electrolytes and were therefore unable to comment upon the rate of change in K⁺ or Ca⁺⁺ during HD. All our patients received acetate HD. Use of bicarbonate as dialysate buffer has been shown to reduce the risk of arrhythmias in some [[10]], but not all [[7]], [[28]] studies. The cardiac status assessment by echocardiography showed that the LVPW thickness was greater and LVEF lower in patients with arrhythmias, but the difference failed to reach statistical significance, probably because of small number of patients in the two groups.

There was an excess of patients with ischemic heart disease amongst those who developed ventricular arrhythmias during HD, suggesting that this could be a predisposing factor. Patients with ischemic heart disease have been shown to have more frequent [[7]], [[15]] and dangerous [[15]] ventricular arrhythmias on HD. This hypothesis is further supported by the finding of silent myocardial ischemia (SMI) on Holter recording in these patients. The role of SMI in the genesis of arrhythmias during HD has not been studied systematically. In one large multicenter study, the frequency of ventricular arrhythmias in patients on HD was found to be similar to that in patients with ischemic heart disease in the absence of any renal insufficiency [[3]]. There has been a lack of agreement on the technique of choice to establish the presence of coronary artery disease in patients without any history of chest pain and a normal resting electrocardiogram. The tests that have been used include exercise testing, radioisotope imaging and ambulatory electocardiographic ST segment monitoring. Abundant clinical and laboratory evidence is now available to show that ST segment depression in EKG in patients with coronary artery disease denotes myocardial ischemia [33-34]. Moreover, Holter monitoring is particularly well-suited to record these transient changes during HD.

Table 5. An overview of literature on the frequency and risk factors of cardiac arrhythmias during HD

Author	Mean age (years)	Duration of dialysis (months)	Ventricular	Supraventricular	Effect of HD	Predisposing factors
Grippo emodialysi [[3]]	55	76	76%	69%	Yes	Age, LV dysfunction
Kimura et al. [[8]]	51	41	77%	68%	Yes	Inceased Ca-P product Impaired cardiac performance
Erem et al.[[9]]	38	12	90%	80%	Yes	Duration of dialysis, acetate
Fantuzzi et al.[[10]]	64	89	94%	NS	Yes	Acetate
Ramirez et al.[[15]]	55	36	33%	32%	Yes	PTH
Abe et al.[[16]]	55	51	12%	7%	Yes	Male sex, age, ischemic heart disease
Weber et al.([[17]]	42	27	33%	0	No	Hypokalemia
Kyriakides et al.[[22]]	41	32	25%	88%	No
Zuber et al.([[25]]	37–82	40	25%	12%	Yes	HD induced ischemia
Nishimura et al.[[26]]	53	55	13.6%	13.6%	Yes	High dialysate calcium
Kayatas et al.([[27]]	45	38	10%	NS	No	Age

Although clinically evident ischemic heart disease was present in only 13% patients, silent myocardial ischemia as evidenced by ST changes on Holter recording was documented in 45% cases. As expected, the proportion of patients with diabetes mellitus and hypertension was higher in this group. Moreover, 61% of patients exhibiting these changes were noted to have ventricular arrhythmias. The frequency of both arrhythmias and ST segment changes was maximum in the last hour of HD. This finding reaffirms the role of coronary ischemia as one of the major factors in the genesis of ventricular arrhythmias in patients on dialysis. In an earlier study, Zuber et al.[[25]] noted ST changes suggestive of SMI in 25% of their patients, half of whom were asymptomatic. Kremastinos et al.[[18]] in a study of 45 patients showed a 15.5% incidence of ST segment changes. However, they could not demonstrate coronary artery disease in any of these cases on angiography. They concluded that ST changes could have been produced by coronary artery spasm secondary to changes in serum K⁺ and Mg⁺⁺ levels. However, in view of the high incidence of coronary artery disease in dialysis patients, total lack of any coronary lesions is difficult to explain. In the present study 29.4% of patients showing ST changes had either angina or a previous myocardial infarction. In contrast, these were present in only 4.8% of patients without ST changes. These findings are similar to those of Abe et al.[[16]] who noted a significantly higher frequency of patients with angina pectoris and those developing chest discomfort during HD amongst those developing HD-related ST changes.

We conclude that cardiac arrhythmias and silent myocardial ischemia are seen frequently amongst CRF patients on HD. High grade arrhythmias were seen in 13%. The simultaneous occurrence of arrhythmias and ST changes towards the end of dialysis points towards a possible etiologic relationship between the two. However, further investigations including coronary angiography are necessary in a larger number of patients to clarify the mechanism of ST changes and conclusively establish the relationship between their occurrence and the presence of CAD.