Electrocardiographic abnormalities in patients receiving hemodialysis

Electrocardiographic abnormalities in patients receiving hemodialysis Shinichi Abe, MD, PhD, a Mamoru Yoshizawa, MD, PhD, b Nagako Nakanishi, MD, PhD, c Tomoko Yazawa, MD, PhD, c Kuninobu Yokota, MD, PhD, d Masasada Honda, MD, PhD, e and Graeme Sloman, AM, ED, BSC, FRCP(LOND) f Tokyo, Urawa, and Kamakura, Japan, and

Melbourne, Australia

We assessed standard 12-lead and Holter electrocardiographic (ECG) abnormalities in maintenance hemodialysis (HD) patients. Of 221 outpatients receiving HD, 143 (65%) had ECG abnormalities. Rates were higher in male, elderly, hypertensive, and diabetic patients than in female, younger, normotensive, and nondiabetic patients. The prevalence of ECG changes correlated inversely with HD duration. Serial ECGs were compared in 87 patients whose average HD duration was 7.5 ± 2.5 years. Thirty-four patients (39%) showed normal ECGs throughout, 27 (31%) relatively stable abnormalities, 22 (25%) worsening, and 4 (5%) reversion to normal. Age, hypertension, and diabetes are factors related to abnormal ECG findings. Among the 142 Holter recordings from 72 patients, 70 (97%) were basically in sinus rhythm, and 2 (3%) were in atrial fibrillation. The average frequency of supraventricular premature contractions (SVPCs) was 1597 ± 9725 per 24 hours, and that of ventricular premature contractions (VPCs), 556 ± 1415. VPCs were multifocal in 9%, in runs in 25%, and early in 1%. In 29 (40%) of recordings, VPCs appeared mainly during and for several hours after HD. ST-T changes were seen in 43 (60%). In 11, ST depression occurred during and a few hours after HD. Patients receiving HD showed diverse ECG abnormalities. Holter ECGs revealed a high incidence of arrhythmias and ST-T changes, which frequently appeared in relation to HD timing. (Am Heart J 1996;131:1137-44.)

It is well known t h a t functional, as well as organic, cardiac problems are a frequent occurrence in patients receiving hemodialysis (HD) for end-stage renal failure. 1,2 According to a number of recent reports, half of patients receiving chronic HD die of cardiovascular disease. 3-5 Reported rates of sudden death in these patients range from 1.4% to 25%. 4, 6 From athe Department of Internal Medicine, Keio University; byoshizawa Clinic; CMiyoshiRenal Clinic; dthe 3rd Department of Internal Medicine, Jikei Medical University; eDepartment of Cardiology, St. Theresa Hospital; fDepartment of Cardiology, Epworth Hospital. Received for publication March 21, 1995; accepted Sept. 11, 1995. Reprint requests: Shinichi Abe, MD, Department ofInternal Medicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 160, Japan. Copyright © 1996 by Mosby-Year Book, Inc. 0002-8703/96/$5.00 + 0 4/1/70738

The Japanese Society for Dialysis Therapy reported in 1991 t h a t -40% of deaths in patients receiving HD were of cardiac origin. 7 The main causes were congestive heart failure, coronary artery disease, and sudden death as a result ofhyperkalemia or arrhythmia. The number of patients with end-stage renal failure (ESRF) receiving maintenance HD has increased over the past few decades in most countries. The numbers of patients who are elderly and have diabetes being given maintenance HD have increased remarkably, 2 and cardiac problems are common in these patients. As the majority of HD patients are cared for at outpatient HD centers, only a small number of those who have cardiac symptoms undergo thorough investigation. However, HD patients have a wide variety of electrocardiographic (ECG) abnormalities and, in certain instances, HD itself seems to be a cause of ECG changes and arrhythmias.1, 4, 5 Arrhythmias are often observed after the start of HD. Therefore, this study was conducted to assess ECG abnormalities in patients receiving maintenance HD with standard 12-lead and Holter ECGs. METHODS

Three separate studies were done. First, standard 12lead ECG data from 221 outpatients receiving HD were analyzed for abnormalities. These results were compared with those of chronic renal failure (CRF) patients not yet receiving HD and those of the normal population. The second study involved follow-up of HD patients for >5 years with standard 12-lead ECGs. All 87 patients who had been receiving HD for >5 years were assessed in this study. ECG data were obtained from these patients once a month, and the findings were evaluated for comparison of serial changes. The third study was based on 24-hour Holter ECG findings during and after HD. Among the 221 patients studied, 72 who had arrhythmias (56) or chest discomfort (16) were selected for Holter ECG evaluation. Twenty-four hour 1137

June 1996 American Heart Journal

1138 Abe et al.

Table I. Background data No. of patients Male Female Average age (yr) Average duration of HD (yr) Cause of ESRF Chronic glomerulonephritis Diabetes mellitus Nephrosclerosis Gout Autosomal dominant polycystic kidney disease Multiple myeloma Obstructive uropathy Tuberculosis No. of Holter ECGs recorded Cardiac complications Angina pectoris Old myocardial infarction Valvular disease Congenital (ASD, VSD) Cardiomyopathy Sick sinus syndrome Medication Digoxin Antiarrhythmic medication

221 142 79 55 _+ 14 (16 to 86 yr) 4.3 +_ 2.4 (1 mo to 11 yr)

Table II.ECG findings in patients receiving HD and the two control groups

WNL

HD patients

Non-HD CRF patients

Bank employees

(n = 2 2 1 )

(n = 70)

(n = 1055)

78 (35.3%) 41 (58.6%) 1005 (95.3%) I

133 49 15 11 9

(60%) (22%) (7%) (5%) (4%)

2 (1%) 1 (0.5%) 1 (0.5%) 142 (72 patients) 11 3 3 2 1 1 8 7

ASD, Atrial septal defect; VSD, ventricular septal defect.

Holter recording was begun just before the start of HD to observe when arrhythmias appeared most frequently. To identify factors possibly related to the occurrence of arrhythmia, body weight, blood pressure, and serum electrolytes, urea nitrogen, and creatinine levels were checked at the start and on completion of HD. Parathyroid hormone (PTH-C) level was also checked at the start of HD. Patient background data, presented in Table I, did not differ significantly among the three study groups. All data on HD patients presented herein were collected from two private outpatient HD centers in the Tokyo metropolitan area. The data from non-HD patients with chronic renal failure were collected at two university hospitals. The normal controls were 1055 company employees who underwent a n n u a l h e a l t h - s c r e e n i n g e x a m i n a t i o n s . For statistical analysis, the chi-square test and t test u s e d . A v a l u e o f p < 0.05 w a s c o n s i d e r e d s i g n i f i c a n t . D a t a a r e p r e s e n t e d a s m e a n v a l u e s - 1 SD. were

RESULTS

The results of the standard 12-lead ECGs of 221 outpatients receiving HD are shown in Table II. One hundred forty-three patients (65%) had ECG abnormalities, not including sinus tachycardia, sinus bradycardia, sinus arrhythmia, and cardiac rotation. The rate of left ventricular hypertrophy (LVH), either alone or with ischemia, was the highest, followed by ventricular and supraventricular pre-

I

I

Abnormal E C G t 143 (64.7%) 29 (41.4%) LVH 41(18.5%) 9 (12.9%) LVH + ischemia 22 (10.0%) 4 (5.7%) VPC 26 (11.8%) 3 (4.3%) Myocardial ischemia 16 (7.2%) 1 (1.4%) SVPC 15 (6.8%) 1 (1.4%) Nonspecific ST-T 13 (5.9%) 2 (2.9%) change AV block 1 degree 12 (5.4%) 1 (1.4%) af 12 (5.4%) 1 (1.4%) LAD ii (5.0%) -CRBBB 7 (3.2%) 3 (4.3%) LA overloading 6 (2.7%) -Hyperkalemia 6 (2.7%) -Old MI 3 (1.4%) -IRBBB 4 (1.8%) 4 (5.7%) Pacemaker r h y t h m 2 (0.9%) -Junctional r h y t h m " 1 (0.4%) -WPW 1 (0.4%) -S-A block 1 (0.4%) -Bifascicular block 1 (0.4%) -RVH 1 (0.4%) --

I

50 (4.7%) 7 (0.7%) -10 (0.9%) 6 (0.6%) 3 (0.3%) 3 (0.3%) ----

1 (0.1%) ---

-19 (1.8%) -1 (0.1%) --

----

CRBBB, Complete right bundle branch block; IRBBB, incomplete right bundle branch block; WPW,Wolff-Parkinson-White syndrome; WNL, within limits. *p < 0.001. ~Two or more abnormalities were noted on some ECG recordings. normal

mature contractions (VPCs, SVPCs), myocardial ischemia, nonspecific ST-T changes, and a variety of other ECG findings. These results were compared with those obtained from the 1055 bank employees and 70 CRF patients (serum creatinine 3.0 to 10.0 mg/dl) not receiving HD, who served as the normal and CRF control groups, respectively. The numbers of male and female normal controls were nearly equal, and their age distribution ranged from the 20s through the 60s. The rate of all ECG abnormalities among this group was only 4.7%. The age range for the non-HD CRF patients was 33 to 83 years (60 ± 13 years); there were 42 men and 28 women; and 41% of these patients had ECG abnormalities. As expected, HD patients had a far higher prevalence of ECG abnormalities t h a n did either of these control groups. The prevalence of ECG abnormalities in the non-HD CRF patients was between those of the normal control and HD groups. Among the HD patients, men had a higher rate of ECG changes t h a n did women (68% vs 53%), and the

Volume 131, Number 6

Abe et al.

American Heart Journal

Total

, p < 0.05 ** p < 0.01 **=~p < 0.001

Sex male female

WNL

LVH

Age N40 41 ---50

I ]

51N6o

ischemia

i]* I

61 ~ 7 0 71~ (yrs) Duration of HD

~2 3~4 5~6 7---8 9-~10 11~ (yrs) Hypertensives Normotensives

1139

VPC

I

SVPC

I I

,

non- specific STT change

I 1

AV block 1

,

]** ~

total hypertensive B B diabetic

LAD

Diabetics Non- diabetics

I

2'0 io

]***

go 8o 10o

Fig. 1. Percentages of HD patients with ECG abnormal-

ities. frequency of ECG abnormalities tended to increase with age. However, there was an apparent inverse relation between the prevalence of ECG changes and the duration of HD (Fig. 1). ECG abnormalities were seen in 70% of patients with hypertension and 91% of those with diabetes. These rates are statistically higher than those of the patients without hypertension and diabetes. Fig. 2 shows comparisons of the percentages of patients with hypertension and diabetes with ECG abnorrealities versus the rate for all HD patients. Both groups tended to have higher rates of left ventricular hypertrophy (LVH), myocardial ischemia, and various arrhythmias. The LVH, ischemia, and other abnormalities were especially common in diabetics. In the second study, 87 patients (62 men, 25 women) who had been receiving HD for an average of 7.5 _+ 2.5 years, were evaluated. The various ECG abnormalities observed during follow-up are shown in Table III. To summarize these findings, 34 (39%) had had normal ECGs for the entire duration of liD, 27 (31%) had relatively stable abnormalities, and 6 of these patients actually showed improvement of ST-T changes, whereas 22 (25%) showed worsening,

af 10

I

20

I

30

I

40

(%)

Fig. 2. Comparison of ECG abnormalities among patients with hypertension, diabetes, and total HD. WNL, Within normal limits; LAD, left axis deviation; af, atrial fibrillation.

and 4 patients' ECGs reverted to normal (Table IV). In comparison with patients whose ECGs were normal throughout, those with stable or worsening abnormalities were older and more likely to be diabetic or hypertensive or both, as shown in Table IV. In fact, as expected, none of the diabetics receiving long-term HD had normal ECG findings. By the end of follow-up, five patients had died, two of cerebral infarction (CI), and one each ofintracerebral hemorrhage (ICH), myocardial infarction (MI), and multiple myeloma (MM)o The average age at death was 75 _+ 16 years (44 to 92 years). Of these five patients, four were hypertensive, and three diabetic. Two of these patients (CI, ICH) had had stable ECG abnormalities, whereas the other two (CI, MM) had shown worsening of ECG findings before death. The patient who died of MI had had normal ECGs throughout the HD period and had never been hypertensive or diabetic. An illustrative ECG is shown in Fig. 3. This 63-year-old man developed ESRF because of autoso-

1140 Abe et

June 1996 American Hea~ Journal

at.

Table III. ECG changes during follow-up :! !

I: l?lli:il~:!li::!!:~.~ !

...... L III~I ~ ;

..: ....

~ii!i:ililqlili!fi!liiiii~i:iiiiii!iii!

! '] .....

: ........

;--.:

......

1-~--!-

Initial W N L (n = 49)

-i-:-i --r-

.

i:.

-

.

_!_~ i i i ! *

. . . . .

T--

Latest

:_._4_,

I

;- - I -

---'-3--"

WNL ST depression LVH LVH + ischemia

Ischemia

.... L*,,-.t---i----i _ . L _ : _ A . ' ! I : :

..............

II

L V H + ischemia (n = 6)

LVH + ischemia

6

L V H (n = 4)

LVH LVH + ischemia

3 1

LAD (n = 5)

LAD LAD + ST depression Bifascicular block

3 1 1

L A overload (n = 5)

LA overload L A overload + L V H L A overload + ischemia + AV block 1 degree LVH + ischemia

2 1 1

WNL

Ischemia

1 3

WNL ST d e p r e s s i o n

2 2

AV block i degree AV block 3 degree

3 1

SVPC (n = 2)

WNL SVPC

1 1

C R B B B (n = 2)

CRBBB

2

I R B B B (n = 1)

IRBBB

1

Old MI (n = 1)

Old MI

1

. I i , : 1 1 ! t

L

~

i

I

I

I

i,

i

-'/---T-I---T-I-T--, -! ...... ::- " ....... !--]---

In

iii ::il:i:it~;:iiff~l~i!if

~--

illiii ii!:

Ischemia (n = 4) ==,L.,. ...... l..I ........... . :::iii:il:i:i!:iiii!~i i.!i! ~gii!iiiiili!! t t :!!':l:i'l!:!F,l.:

j :i:

:L.

i!:: i!: !;:i :! ~ J : : : L : L : " :1 "

: :7::,:':=1

::::1..:

.:;

:::F:::I::::I

: .!.:-:

: 1": :

i

i

:

~ q,~-rP-F i

i

.

ST d e p r e s s i o n (n = 4)

AV block 1 degree (n = 4)

,-:

i

i

[

i~i!li!sli::

1

:

Vl ii~!iii!!t.i-!}:iii

34 7 4 3 t

ii:::

Fig. 3. Bifascicular block appeared in this patient with autosomal dominant polycystic kidney disease after several years of HD. Left column shows October 1991 E C G ; right column shows November 1991 E C G . mal dominant polycystic kidney disease and had been receiving HD for 8 years when the tracing was obtained. After the development ofbifascicular block, without subjective symptoms, an echocardiogram was done but revealed no significant findings. The bifascicular block was unchanged as of the most re: cent follow-up in February of 1994. The ECG of a 49-year-old man who was given HD for ESRF, caused by chronic glomerulonephritis, is shown in Fig. 4. He initially had inverted T waves that normalized within a month of initiating erythropoietin treatment. The erythropoietin treatment produced a concomitant increase in serum hemoglobin from 6.5 guddl to 9 . 0 g m / d l . Serial ECG changes in a 57-year-old male diabetic HD patient are shown in Fig. 5. A t the initiation of

WNL, Within normal limits.

HD, he had first-degree atrioventricular (A-V) block. This continued for nearly 4 years, until the patient noticed a slow pulse rate. As shown in the figure, second-degree A-V block had developed. Within a month, he had complete A-V block with no subjective symptoms. Artificial pacing was instituted. The third study focused on Holter ECG findings. Among the 221 patients receiving HD, 72 who had had arrhythmias (56) or chest discomfort (16) were selected for evaluation. Arrhythmias had been identified on physical examination, either by standard 12-lead ECG or as a consequence of subjective symptoms. Of the 16 patients with chest discomfort, 11 had angina pectoris, whereas the other 5 had atypical chest pains. The results of 142 recordings from these 72 patients are shown in Table V.

Volume 1 3 1 , N u m b e r ' 6 American Heart Journal

Ab8 et al.

::-_i-~i~::.ii~ 1! i~i-i~!~:i~/l::-::::]~ii]i :-:.-:

'==:1:'=:- ........ : .......

~, . . . . . . . .

[iiiiii:-~iii;![:~

~

,

;

:=,

: : =:: :

1141

(Dec. 1993)

li=i:-==

.......

T ......

T -=-~

. . . . . . . . . .

~:..~..L-r-~:~. ::::::~:

H !!il J l i i H i ~ i l i ! ~ - ~ l i i "

:

::1

(Feb. 1994)

-ii-

V4

ii

.:?:i::i~:!i~]i ....... = ....

-~_: ~__-

. ~

:i-"i ~' i'iii!:::i 'ii!ii::iii~i!.ii!i

]:

(March 1994)

:~

j~_~.,

,L__../-~

: : : : : : . ;i;f :;:;i,:i;:;::ii;:i;:i;:;iiiii:i:i ¸ ::~:-::::1::: ::::;i;!i

~i:i ~i ii!i!iiiii

!iiil~iiiiii:!ii~

~:

ii

:

! .......

:=i

i

(March 1994)

:

: ~

i

Fig. 5. Progression from first-degree to complete A-V block in patient with diabetes receiving HD.

:

:.

I:::::::

::::: ::

:: : ; . : : : : : :

:iii-i:!ii:.::~:~::=:::-ii:! !i!~i!i

V6

• :i:!

['

; ~F~[i:\i'!~:l!=:

~?:::I:F:{::

..... ;. =. ::~i~:iii!i:!-i~it :i:.::~i~

'" ±" : : : " r .... :'" :1 : q : : : H : : ' : : : ; : ' = '

..........

Fig. 4. Improvement in ST-T changes in a patient with chronic glomerulonephritis. Left column shows February 1991 ECG; right column shows March 1991 ECG. Forty-two patients were men and 30 were women, and their average age was 58 ± 13 years. Seventy patients (97%) had basically sinus rhythm, including 3 (4%) with transient atrial fibrillation, and 2 (3%) had atrial fibrillation. SVPCs were observed in 68 (94%) patients and VPCs in 62 (86%). The average frequency of SVPCs was 1597 ± 9725 (1 to 99,671) per 24-hour period, and t h a t of VPCs was 556 ± 1415 (1 to 8336). VPCs were multifocal in 9%, in runs in 25%, and early in 1%. Therefore, 35% of these recordings revealed VPCs more severe than Lown's grade 3. ST-T changes were seen in 43 patients (60%); ST depression in 31, elevation in 8, and T inversion in 4. Among the patients who had arrhythmias, only 7 needed antiarrhythmic medication (mexiletine or disopyramide). The remainder were closely moni-

tored without medication. Forty patients had repeated Holter ECG recordings. In 10 (25%) of 40 patients with SVPCs and 14 (35%) of 40 with VPCs, the frequency of arrhythmias decreased by ->50%. In the remaining patients, none of whom had subjective symptoms, the arrhythmias were unchanged. Two patterns were identified on 24-hour Holter ECGs (Fig. 6). In 19 (27%) of the Holter recordings with SVPCs and 29 (40%) of those with VPCs, arrhythmias were observed to be related to the timing of liD. The relations between arrhythmia pattern and various clinical parameters were assessed in patients who had >100 VPCs per 24 hours (Table VI). The average change in serum K was significantly larger in the HD-related arrhythmia group. Variables such as average changes in body weight and serum Ca, Mg, urea nitrogen, and serum creatinine levels were somewhat higher, whereas the serum K levels were somewhat lower, at completion, and serum Ca levels were lower and PTH-C levels were somewhat higher at the start of HD, in this group. The differences were not, however, statistically significant. As to other variables, there were essentially no differences between the two groups. Among the 43 patients who had ST-T changes, 11 (26%) had changes related to the timing of HD, as shown in Fig. 7. The percentages of chest discomfort, a diagnosis of angina pectoris, and hypertension

June 1996 American Heart Journal

1142 Abe et al.

Table IV. ECG changes during follow-up

WNL throughout Stable abnormalities Worsening Reverted to normal Total

No. of patients

Sex ratio (M / F)

34(39%) 27 (31%) 22 (25%) 4 (5%) 87

1.8 3.5 3.4 1.0 2.5

Average age 44.5-11.67,1 58.5 -+ 15.4 56.6 - 11.4 43.3 -+ 5.9 51.8 -+ 14.4

,

Hypertensive

Diabetic

15 (44%) ~ t 1 22 (81%) 19 (86%) t 1 (25%) 57 (66%)

0(0%)7, 7 9 (33%)~ | , 7 (32%) l 0 (0%) 16 (18%)

Data reported as mean ± 1 SD. *p < 0.001. tP < 0.01.

Table V. Comparison of Holter and standard ECG fmdings Holter ECG

Standard ECG

Table VI. Relations between arrhythmia pattern and clin-

ical parameters in patients with >100 VPCs per 24 hours HD related

Basic r h y t h m Sinus r h y t h m 70 (97%) Transient af 3 (4%) af 2 (3%) SVPC 68 (94%) Mean _+ 1 SD 1597 +_ 9725 Range 1 - 99,671 VPC 62 (86%) Mean _+ 1 SD 556 - 1415 Range 1 - 8336 Lown's grading for VPCs 0 14% 1 45% 2 7% 3 9%| 4 25%~ 35% 5 l%J ST-T changes 43 (60%) ST depression 31 (43%) ST elevation 8 (11%) Inverted T 4 (6%)

70 (97%) 3 (4%) 2 (3%) 10 (14%)

20 (28%)

17 (24%)

were significantly higher in these patients than in those whose ST-T changes had no relation to HD timing (Table VII). DISCUSSION

It is important for a patient receiving maintenance HD to retain adequate cardiac function to have a good quality of life for as long as possible. Although it is not uncommon for patients with ESRF to have a variety of cardiac problems at the initiation of liD, 1 most of these symptoms have subsided by the time routine outpatient HD has been established. Therefore, most patients are not aware of having any cardiac problems, in spite of the variety of ECG abnormalities commonly found in those receiving HD, as shown in Table II. It is reasonable to speculate, comparing the data from the three groups, that the high percentage of abnormal ECGs in patients receiving HD m a y be the

Weight (kg) Systolic pressure (mm Hg) Diastolic pressure (ram Hg) Na (mEq/L) K (mEq/L) K* (mEq/L) Ca (mg/dl) Cat (mg/dl) Mg (mg/dl) Mg* (mg/dl) Urea nitrogen (mg/dl) Urea nitrogent (mg/dl) Creatinine (mg/dl) Creatininet (mg/dl) PTH-C~ (ng/ml)

HD unrelated

2.2 _+ 0.4 18.3 -+ 12.9

1.9 _+ 0.7 22.4 _+ 11.1

5.1 + 4.3

9.0 -+ 8.2

2.5 1.8 3.4 1.5 8.5 0.5 2.1 53.6

_+ 2.3 -+ 0.5 _+ 0.3 +- 0.7 -+ 0.7 _+ 0.1 +_ 0.1 -+ 13.6

2.1 1.3 3.6 1.0 9.2 0.4 2.1 44.7

_+ 1.8 -+ 0.4 (p < 0.05) _+ 0.4 -+ 0.7 -+ 0.6 +- 0.2 -+ 0.1 +- 10.3

82.6 _+ 21.5

69.5 _+ 15.1

7.0 +- 1.8 11.3 -+ 2.8 4.7 -+ 3.4

6.4 _+ 1.7 10.8 +- 2.5 2.8 +- 1.8

Data reported as mean ± 1 SD. Values without superscripts indicate difference between start and completion of HD. *Value at completion of HD. ~Value at start of HD.

result of long-standing CRF with hypertension, anemia, the effects of so-called uremic toxins, and volume overload, rather than the effect of HD itself. As expected, male, elderly, hypertensive, and diabetic patients had higher rates of ECG abnormalities than female, younger, normotensive, and nondiabetic patients, respectively. Contrary to our expectations that those who had been receiving HD for a long time would have more severe ECG changes, possibly as a result of atherosclerosis, Ca deposition, or dialysis amyloidosis, there was an apparent inverse correlation between the prevalence of ECG changes and the duration of HD. In our experience, the onset of severe cardiac abnormalities in a patient receiving long-term HD heralds a decline in the quality of life. These patients are frequently hospitalized for man-

6

Volume 131, Number

<

HD •

+1°to < HD

300 t VPC/hr

1143

Abe et al.

American Heart Journal

2001

ST

1oo1

Level -1,0

(mm)

-2.0

15:00

20:00 0:00

6:00

15': 00

12:00

HD

ST Level _1.0.= ~ (mm) -2,0

1001 15:00

3:'00

9:'00

+1.~1 ( HD )

VPC/hr 200] i

21 ': 00

i

20:00 0:00

r

t

9 :'00

6:00

12:00

6. Arrhythmia in relation to HD timing. Two patterns are shown; (1) arrhythmias appearing during and 5 to 6 hours after HD (lower, 78-year-old woman with chronic glomerulonephritis; HD duration of 4 years) and (2) arrhythmias unrelated to HD timing (upper, 73-year-old man with chronic glomerulonephritis; HD duration of 3 years). Fig.

agement of congestive heart failure or coronary artery disease or both. Diabetics have high rates of LVH, myocardial ischemia, and various arrhythmias. It is common for diabetics to have so-called diabetic cardiomyopathy. Recently an investigation was reported focusing on microangiopathy as the pathogenesis of cardiomyopathy, s The results of this study, conducted from the viewpoint of microvascular structure within the h u m a n heart, indicate t h a t the capillaries of diabetics develop irregular narrowing, tapering, occlusion, and microaneurysmal configurations. These microvascular abnormalities were speculated to have played a major role in the pathogenesis of the cardiomyopathy, which produced the abnormalities revealed by ECG in diabetics. These findings support our data showing t h a t diabetics have a variety of ECG abnormalities t h a t persisted throughout the follow-up period. The results of Holter ECG observation indicate that patients receiving HD have extremely high arr h y t h m i a rates, especially VPCs classified as 3, 4, and 5 of Lown's grading. The same results have been reported previously. 4, 5 If such arrhythmias are a cause of sudden death in HD patients, careful follow-up or administration of antiarrhythmic drugs or both should be considered. Chazan 6 reported t h a t 25 of 31 in-center cardiac arrests were associated with ventricular fibrillation. We had six patients who died while bathing or sleeping. As none was being monitored at the time of death, we can only speculate t h a t some of these deaths might have been the result of an arrhythmia. No autopsies were performed.

15'; 00

~'~

~

•

15': 00

21 ': 00

~ =

=

3 :'00

9 :'00

Fig. 7.

ST-T changes in relation to HD timing. Again, two patterns can be discerned: (1) ST depression appearing during and 2 to 3 hours after HD (lower, 75-year-old man with chronic glomerulonephritis; HD duration of I month) and (2) ST depression unrelated to HD timing (upper, 65year-old woman with nephrosc]erosis; HD duration of 10 years).

TableVII. Relations between ST-T changes on Holter ECG and timing of HD

No. of patients Male/female Average age (yr) Chest discomfort

Related

Not related

11 (26%) 4/7 61 -- 10 6 (55%)

32 (74%) 18/14 56 _+16 5 (16%)

I Ischemic change on standard ECG Clinical diagnosis of angina pectoris

Hypertensive

4 (36%)

I *

5 (45%) I

11 (34%) 3 (9%) I

11 (100%)

22 (69%)

I Diabetic

3 (27%)

I *

7 (22%)

*p < 0.05. ~p < 0.01.

Only seven of our patients needed antiarrhythmic medications for control of subjective complaints. In contrast to patients with MI, indications for administration of these drugs are limited in HD patients. Because the pathogenesis of arrhythmias and longterm effects of such drugs have not been adequately assessed in HD patients, minimizing or removing factors possibly inducing arrhythmias, such as those described below, could be the most appropriate form of treatment. In the 40% of patients who had VPCs, they appeared soon after the start and disappeared 5 or 6 hours after the completion of HD. Even when the well-known diurnal variation in ventricular arrhyth-

1144 Abe et al.

mias 9 is taken into consideration, it is clear that the VPCs shown in Fig. 6 are related to HD. Data from previous studies suggest that changes in serum 1° and intracellular n' 12 K, as well as those in serum Ca 1, 13 and Mg, decreased circulating blood volume, 1 rapid correction of metabolic acidosis, n, 13 and increases in parathyroid hormone, 13, 14 serum free fatty acids 15 and age, 1~ may contribute to the occurrence of arrhythmias. Our results indicate that serum K changes during HD play a role in triggering arrhythmias. Rapid decreases in circulating blood volume may also contribute to the onset of arrhythmias. Five patients who had VPCs in relation to HD showed few VPCs when the average amount of fluid removed by HD was reduced from 2.61 to 2.01 L. Although it was not possible to identify factors related to the occurrence of arrhythmia, a combination of these changes appears to trigger arrhythmias in patients with latent cardiac problems, 17 resulting in the occurrence of these arrhythmias in relation to the timing of HD. In this investigation, we only touched on the issue of ST-T changes, which are also related to the timing of liD in a significant proportion of patients. This will be the focus of future studies examining how ST-T changes are related to clinical symptoms, blood chemistry, exercise ECGs, echocardiographic findings, and myocardial scintigraphy. Further investigation, aimed at clarifying the mechanisms underlying arrhythmias and ST-T changes in patients receiving HD is needed to improve both the quality of life and prognosis of these patients. REFERENCES

1. Harnett JD, Parfrey PS. Left ventricular dyshmction in dialysis subjects. In: Henrich WL, ed. Principles and practice of dialysis. Baltimore: Williams & Wilkins, 1994:170-80.

June 1996 American Heart Journal

2. Rostand SG, Rutsky EA. Coronary artery disease in end stage renal disease. In: Henrich WL, ed. Principles and practice of dialysis. Baltimore: Williams & Wilkins, 1994:181-95. 3. United States Renal Data System. 1993 Annual data report: causes of death. Am J Kidney Dis 1993;22(suppl 2):53-7. 4. Gruppo Emodialisi e Patologie Cardiovascolari. Multicenter, cross-sectional study of ventricular arrhythmias in chronically haemodialysed patients. Lancet 1988;2:305-8. 5. KimuraK, TabeiK, AsanoY, HosodaS. Cardiacarrhythmiasinhemodialysis patients: a study of incidence and contributory factors. Nephran 1989;53:201-7. 6. Chazan JA. Sudden death in patients with chronic renal failure on hemodialysis. Dial Transplant 1987;16:447-8. 7. Japanese Society for Dialysis Therapy Committee report. The chronic dialysis patients registry in Japan--199L J Jpn Soc Dial Ther 1993; 26:17-30. 8. Yokota K, Furusato M. The abnormalities of microvascular structure of the h u m a n diabetic heart. In: Nagano M, Dhalla NS, eds. The diabetic heart. New York: Raven Press, 1991;103-13. 9. Raeder EA, Hohnloser SH, Graboys TB, Podrid PJ, Lampert S, Lown B. Spontaneous variability and eircadian distribution of ectopic activity in patients with malignant ventricular arrhythmia. J Am Coll Cardiol 1988;12:656-61. 10. Morrison G, Michelson EL, Brown S, Morganroth J. Mechanism and prevention of cardiac arrhythmias in chronic hemodialysis patients. Kidney Int 1980;17:811-19. 11. Fantuzzi S, Caico S, Amatruda O, Cervini P, Abu-Turky H, Baratelli L, et al. Hemodialysis-associated cardiac arrhythmias: a low risk with bicarbonate? Nephron 1991;58:196-200. 12. Rombola G, Colussi G, Elisabetta DeFerrari M, Frontini A, Minetti L. Cardiac arrhythmias and electrolyte changes during haemodialysis. Nephrol Dial Transplant 1992;7:318-22. 13. Ramirez G, Brueggemeyer CD, Newton JL. Cardiac arrhythmias on hemodialysis in chronic renal failure patients. Nephron 1984;36:21218. 14. Gafter U, Battler A, Eldar M, Zevin D, Neufeld HN, Levi J. Effect of hyperparathyroidism on cardiac function in patients with end-stage renal disease. Nephron 1985;41:30-3. 15. Suzuki Y, Narita M, Yamazaki N. Effect of L-carnitine on arrhythmias during hemodialysis. Jpn Heart J 1982;23:349-59. 16. Shapira OM, Bar~Khayim Y. ECG changes and cardiac arrhythmias in chronic renal failure patients on hemodialysis. J Electrocardiol 1992;25:273-9. 17. Wizemann V, Kramer W, Thormann J, Kindler M, Schutterle G. Cardiac arrhythmias in patients on maintenance hemodialysis: causes and management. Contrib Nephrol 1986;52:42-53.