MAJOR ARTICLE

Clinical Features and Predictors of Diphtheritic

Cardiomyopathy in Vietnamese Children
Rachel Kneen,1,3 Nguyen Minh Dung,2 Tom Solomon,1,3 Pham Ngoc Giao,2 Christopher M. Parry,1,3
Nguyen Thi Tuyet Hoa,2 Ha Thi Loan,2 Ann Taylor,3 Vo Thi Thien Huong,2 Nguyen Thi Thu Nga,2
Nicholas P. J. Day,1,3 and Nicholas J. White1,3
1
Wellcome Trust Clinical Research Unit, Centre for Tropical Diseases, and 2Centre for Tropical Diseases, Cho Quan Hospital, Ho Chi Minh City,
Vietnam; and 3Centre for Tropical Medicine, Nuffield Department of Clinical Medicine, John Radcliffe Hospital, Oxford, United Kingdom

Background. Despite the availability of antitoxin and antibiotics, the mortality rate for diphtheria remains
high, mostly because of cardiac complications.
Methods. During 1 year, 154 Vietnamese children with diphtheria admitted to a referral hospital were studied
prospectively with clinical examination, including a simple pseudomembrane score, 12-lead and 24-hour electro-
cardiography, measurement of serum cardiac enzyme levels, and estimation of troponin T levels.
Results. Thirteen children had diphtheritic cardiomyopathy on admission, and 19 developed it subsequently.
Twelve children (8%) died. The combination of pseudomembrane score of 12 and bull neck predicted the de-
velopment of diphtheritic cardiomyopathy, with a positive predictive value of 83% and a negative predictive value
of 93%. Administration of 24-hour electrocardiography on admission improved the ability to predict diphtheritic
cardiomyopathy by 57%. Fatal outcome was best predicted by the combination of myocarditis on admission and
a pseudomembrane score of 12. Of the cardiac enzyme levels measured, an elevated aspartate aminotransferase
level was the best predictor. The presence of troponin T identified additional children with subclinical cardiac
damage.
Conclusions. The development of diphtheritic cardiomyopathy can be predicted by means of simple measures.

Diphtheria remains an important cause of child mor-
tality in developing countries. The mortality rate is still
∼10% and has changed little over the past 100 years
[1]. Acute mortality is due to toxin-mediated diphthe-
ritic cardiomyopathy, suffocation by the pseudomem-
brane, disseminated intravascular coagulation, and
renal failure [2, 3]. The incidence of diphtheritic car-
diomyopathy following diphtheria is 10%–20%, and
the associated mortality is ∼50% [4]. Clinical signs of
diphtheritic cardiomyopathy become apparent by the
end of week 2 of infection but, in severe cases, may be
a presenting feature [5]. Most of the large series de-
scribing the clinical and electrocardiographic features
of diphtheria were reported 160 years ago, before the
availability of modern electrocardiographic and bio-
chemical measurements [6–9]. These studies found that
the development of severe conduction defects on the
12-lead electrocardiograph were associated with a poor
prognosis. We have recently reported on the prognostic
utility of 24-h electrocardiography [5] and have shown
that, in some cases of diphtheritic cardiomyopathy, in-
tervention with temporary cardiac pacemaker may im-
prove the outcome [10]. The ability to predict from
simple and readily available measures whether myo-
carditis will develop would aid in triage and clinical
management. We therefore conducted a prospective
study of newer diagnostic tools to assess clinical and
laboratory findings that might predict the development
of diphtheritic cardiomyopathy and poor outcome.

PATIENTS AND METHODS

Received 20 February 2004; accepted 29 June 2004; electronically published 8
November 2004.
Reprints or correspondence: Dr. Rachel Kneen, Roald Dahl EEG Unit, Royal
Liverpool Children’s NHS Trust, Alder Hey, Liverpool L12 2AP, United Kingdom
(rachel.kneen@blueyonder.co.uk).
Clinical Infectious Diseases 2004; 39:1591–8
 2004 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2004/3911-0006$15.00
Patients. The study was conducted at the Centre for
Tropical Diseases (a referral center for patients with
diphtheria from Southern Vietnam), Ho Chi Minh City,
Vietnam. The study was approved by the center’s Sci-
entific and Ethical Committee, and informed consent
was obtained. Consecutive children (age, !17 years)

Diphtheritic Cardiomyopathy in Children • CID 2004:39 (1 December) • 1591

Downloaded from https://academic.oup.com/cid/article-abstract/39/11/1591/463517

by guest
on 02 August 2018
 with a diagnosis of diphtheria were admitted prospectively to
the study during the period of April 1995 through March 1996.
Diphtheria was diagnosed clinically either if the patient had a
febrile illness with a characteristic adherent pseudomembrane
visible in the nasopharynx or if the patient presented later (after
pseudomembrane clearance) with a history of recent severe sore
throat and signs of cardiomyopathy. We defined symptomatic
diphtheritic cardiomyopathy in patients who developed symp-
toms of, and examination findings consistent with, heart failure
(see Results) and abnormal findings on 12-lead electrocar-
diography. Children with no symptoms of heart failure, but
with either clinically detected rhythm disturbances or abnormal
findings on 12-lead electrocardiography, were defined as having
asymptomatic diphtheritic cardiomyopathy. Patients were given
a severity score defined as follows: “mild,” local symptoms only;
“moderate,” patient is systemically unwell with a “toxic” facial
appearance; “severe,” patient is bed-bound, is unable to drink,
has difficulty breathing, or has alteration in mental status (table
1). We also prospectively devised a score that estimated the
amount of pseudomembrane visible in the nasopharynx on
admission to see whether this was a better predictor (table 1).
Patients were examined daily and again 1 month after discharge.
Microbiological investigations. Four sterile swab speci-
mens were taken on admission (1 from each tonsil and 1 from
each nostril). An additional set of swab specimens was also
obtained on days 2, 5, 12, 24, and at follow-up. Swabs were
inoculated onto 5% sheep blood agar and Hoyles tellurite agar
with 5% lysed sheep blood (Oxoid) [4]. Plates were incubated
in air at 37C for 48 h. Suspicious colonies were identified as
Clostridium diphtheriae on the basis of the pyrazinamidase and
cystinase test results [11], and identification was confirmed
using a commercial kit (API Coryne). Toxigenicity was deter-
mined by use of the ELEK immunoprecipitation test [4, 11].
Disk antimicrobial susceptibilities were determined as described
previously [4, 11, 12].
Clinical procedures. Blood was drawn on admission for
determination of hematocrit, differential WBC count, platelet
count, plasma biochemistry, and standard cardiac enzyme level
estimation (creatinine phosphokinase, lactate dehydrogenase,
and aspartate aminotransferase [AST] levels). Blood was also
drawn for estimation of cardiac troponin T (cTnT) level at
admission, weekly, at follow-up, and more frequently in severely
ill patients. Samples for cardiac enzyme and cTnT level esti-
mation were centrifuged immediately and were stored at ⫺20C
until further analyses. cTnT levels were measured with the En-
zymun-Test immunoassay (Boehringer Mannheim). The de-
tection limit of the cTnT assay was 0.01–16.6 ng/mL [13]. The
presence of any cTnT was considered to be abnormal [14].
Abnormal levels of cardiac enzymes were defined as follows:
creatinine phosphokinase, 1200 IU/L; lactate dehydrogenase,
1250 IU/L; and AST, 147 IU/L.
1592 • CID 2004:39 (1 December) • Kneen et al.
Downloaded from https://academic.oup.com/cid/article-abstract/39/11/1591/463517
by guest
on 02 August 2018
Table 1. Severity and membrane scores at presentation for 154
children with diphtheria (32 with diphtheritic cardiomyopathy).
No. of patients (no. with
diphtheritic cardiomyopathy)
Children with
fatal cases Survivors
Score (n p 12) (n p 142)
Severity score
Mild 0 (0) 42 (0)
Moderate 0 (0) 78 (7)
Severe 12 (12) 22 (13)
Membrane score
0a 0 (0) 2 (2)
1 0 (0) 29 (2)
2 0 (0) 92 (7)
3 3 (3) 17 (7)
4 9 (9) 2 (2)
NOTE. Severity score: mild, local symptoms only; moderate, patient is
systemically unwell with a “toxic” facial appearance; severe, patient is bed-
bound, is unable to drink, has difficulty breathing, or has alteration in mental
status. Membrane score: 0, membrane cleared before presentation; 1, nose
only or incomplete/follicular coverage of tonsils; 2, confluent coverage of ton-
sils; 3, as above, plus palate and/or pharyngeal wall; 4, as above, plus nose
and/or larynx.
a
Two patients with diphtheritic cardiomyopathy presented after the
membrane had cleared.
Twelve-lead electrocardiography was done at admission and
as clinically indicated. Continuous electrocardiographic mon-
itoring for 24 h was done with a Medilog 4000-II monitor
(Oxford Instruments) on admission and at weekly intervals
depending on the availability of the monitors. Twenty-four–
hour electrocardiographic recordings were analyzed at a later
date by a technician blind to the clinical details. Results were
interpreted with the use of tables of normal values, and QT
interval corrected for heart rate (QTc) was calculated by use of
Bazett’s formula [15].
We used definitions of abnormalities used previously at our
health care center when interpreting the 24-h electrocardio-
graphs [5]. Because it was anticipated that 24-h electrocar-
diography might yield several abnormal findings, a 24-h elec-
trocardiographic abnormality score was devised in which each
different abnormality detected (excluding sinus tachycardia)
counted as 1 point.
Treatment. At admission, patients were treated with 20–
100,000 IU of equine diphtheria antitoxin (Pasteur Institute)
administered intramuscularly in accordance with World Health
Organization guidelines [16]. Patients with severe diphtheria
were treated with benzyl penicillin at (100,000 IU/kg iv q.d. for
14 days). Patients with mild or moderate disease were included
in a treatment trial [4] and received either intramuscular benzyl
penicillin (50,000 IU/kg q.d. for 5 days), followed by oral pen-
icillin V (50 mg/kg q.d. for 5 days), or oral erythromycin ethyl-
succinate (50 mg/kg q.d. for 10 days). Severe tonsillar and
 Table 2. Clinical features of 154 children with diphtheria.

Children with
fatal cases Survivors
Feature (n p 12) (n p 142) OR (95% CI) P
Male sex 5 (42) 84 (59) 0.49 (0.13–1.84) .24
Age, median years (range) 7.5 (1.75–12) 5 (0.75–14) NA .08
Duration of illness, median days (range) 4.5 (3–12) 3 (1–20) NA .017
Rural residence 12 (100) 49 (35) NA !.0001
Sore throat 12 (100) 126 (89) NA .25
Cough 5 (42) 74 (52) 0.66 (0.17–2.45) .16
Malaise 8 (67) 62 (44) 2.58 (0.66–10.75) .22
Nasal discharge 5 (42) 44 (31) 1.59 (0.41–6) .32
Hoarse voice 10 (83) 38 (27) 13.68 (2.63–95.09) .0001
Facial or neck swelling 11 (92) 27 (19) 46.85 (5.8–1012.7) !.0001
Full primary diphtheria toxoid
immunization 0 35 (25) NA .04
Pretreatment with antibiotics 7 (58) 69 (49) 0.68 (0.18–2.52) .7
Received antitoxin before day 3 of illness 0 35 (25) NA .04
Admission temperature, median C (range) 37.5 (37–39.5) 38.2 (37–40.8) NA .01
Bull neck 12 (100) 22 (15) NA !.0001
Bleeding tendency 9 (75) 12 (8) 32.5 (6.7–178.8) !.0001
Stridor requiring tracheotomy 6 (50) 6 (4) 22.67 (4.67–117.15) !.0001
Diphtheritic cardiomyopathy at hospital
admission 10 (83) 3 (2) 231.7 (27.8–2811.5) !.0001
Clostridium diphtheriae isolated 11 (92) 23 (16) 56.9 (16.9–1237.3) !.0001

NOTE. Data are no. (%) of patients, unless otherwise indicated. NA, not applicable.

laryngeal edema was treated with hydrocortisone (2–4 mg/kg RESULTS

q.d. iv for 5–7 days). Tracheotomy was done if edema or la-
ryngeal pseudomembrane caused airway obstruction. Pulmo-
nary edema was treated with furosemide (1-mg/kg iv per dose,
as required), and cardiogenic shock was treated with inotropic
infusions (dopamine, 2.5–10 mg/kg/min iv, or dobutamine, 2.5–
10 mg/kg/min iv). Patients with diphtheritic cardiomyopathy
who developed a severe symptomatic bradyarrhythmia had a
temporary cardiac pacing wire inserted until sinus rhythm was
restored or they died.
Statistical analysis. Normally distributed data were com-
pared by Student’s t test; data that were not normally distrib-
uted were compared by the Mann-Whitney U test. Differences
between proportions were tested using the x2 test with Yates’s
correction or Fisher’s exact test (Statview, version 4.02; Abacus
Concepts) [17]. Variables associated with the development of
diphtheritic cardiomyopathy in univariate analyses were in-
cluded in a stepwise logistic regression to create a model pre-
dictive of diphtheritic cardiomyopathy (SPSS, version 9). Terms
were entered into the model and remained in only if they were
statistically associated with the development of diphtheritic car-
diomyopathy (P ! .05). On the basis of the results of this logistic
regression, the sensitivity, specificity, positive predictive value
(PPV), and negative predictive value (NPV) of the model were
tested. A similar logistic regression was performed to develop
a model predictive of fatal outcome.
During the 1-year study, 154 children with a clinical diagnosis
of diphtheria were admitted to the hospital. The median age
was 5.5 years (range, 0.75–14 years). Eighty-nine (58%) were
boys. Four children (ages 6–13 years) were from the same fam-
ily, including a 13-year-old child with an unrepaired cleft palate.
Thirteen children had diphtheritic cardiomyopathy at hospital
admission, and 19 developed it subsequently. Twelve children
(8%) died between 4 and 18 days (median, 10 days) after onset
of the infection. Eight children (5%) developed diphtheritic
neuropathy as a late complication. A description of their clinical
courses will be the subject of a separate report.
Admission clinical features. Children typically presented
with an acute febrile illness with a severe sore throat, cough,
and malaise. A history of facial or neck swelling or hoarse voice
was more frequent in patients who died (table 2). Those who
received antitoxin early (before day 3 of symptoms) had a better
outcome. All but 2 patients had pseudomembrane visible in
the nasopharynx. Both of these patients presented late with a
history of recent severe sore throat and had diphtheritic car-
diomyopathy (on days 7 and 20 of the illness). Children with
the following characteristics noted during initial examination
were more likely to die: bull neck (figure 1); mucosal, skin, or
nasal bleeding (figure 2); severe airway obstruction requiring
a tracheotomy; a pseudomembrane score of 12; or an initial

Diphtheritic Cardiomyopathy in Children • CID 2004:39 (1 December) • 1593

Downloaded from https://academic.oup.com/cid/article-abstract/39/11/1591/463517

by guest
on 02 August 2018
 Figure 1. A 10-year-old girl with diphtheria presenting with bull Figure 2. A 5-year-old boy with diphtheria presenting with severe
neck on day 4 of illness. nasal bleeding and skin purpura on day 3 of illness.

severity score of “severe” (tables 1 and 2). Forty-nine children fatal outcome (table 4). In a multiple logistic regression model,
(32%) had copious clear or blood-tinged nasal discharge, with the combination of bull neck and pseudomembrane admission
pseudomembrane visible in the nares (figure 3). Twelve of 31 score of 12 gave the best predictor of the development of diph-
patients with a pseudomembrane score of 12 died, compared theritic cardiomyopathy, correctly identifying 10 of 19 patients
with none of those with a lower pseudomembrane score (P ! who developed it and 120 of 122 patients who did not (sen-
.0001). In a multiple logistic regression model, the combination sitivity, 53%; specificity, 98%; PPV, 83%; NPV, 93%; OR, 31.6;
of the presence of diphtheritic cardiomyopathy at hospital ad- 95% CI, 5.9–306; P ! .0001). Among the 20 survivors of diph-
mission and a pseudomembrane score of 12 proved the best theritic cardiomyopathy, symptoms or abnormal clinical car-
predictor of fatal outcome, correctly identifying 10 of 12 fatal diac findings remained present for median of 10 days (range,
cases (sensitivity, 83%; specificity, 100%; PPV, 100%; NPV, 3–35 days).
99%; OR, 232; 95% CI, 35–1150; P ! .0001). The 28 children with symptomatic diphtheritic cardiomy-
Renal complications. Oliguric renal failure developed in opathy typically were tired and listless, looked pale and clammy,
all patients who died. The serum creatinine level on admission and had decreased appetite or vomiting. Frequent findings of
(corrected for age and sex) was more likely to be elevated in physical examinations included tachypnea, hepatomegaly, quiet
fatal cases (16 nonfatal cases vs. 11 fatal cases; OR, 86.6; 95% heart sounds, signs of cardiac enlargement, tachycardia, pal-
CI, 10.3–1916.8; P ! .0001) (table 3). Therefore, elevated serum pable ectopic beats, irregularly irregular heart beat, gallop
creatinine levels at admission predicted fatal outcome with a rhythm, new murmurs, poor peripheral perfusion, widening of
sensitivity of 92%, a specificity of 89%, a PPV of 41%, and an pulse pressure, and in severe cases, hypotension. Patients with
NPV of 99%.
Diphtheritic cardiomyopathy. Thirty-two children (21%)
developed either symptomatic or asymptomatic diphtheritic
cardiomyopathy: 13 (41%) had evidence of diphtheritic car-
diomyopathy at hospital admission, and 19 (59%) developed
diphtheritic cardiomyopathy subsequently. Twenty-eight pa-
tients with diphtheritic cardiomyopathy were symptomatic, but
4 remained asymptomatic (with cardiac rhythm abnormalities
as the only manifestation). Twelve children (37.5%) with diph-
theritic cardiomyopathy died. The median time from onset of
the infection to the manifestations of diphtheritic cardiomy-
opathy was shorter in fatal cases (5 days [range, 3–12 days] vs.
8 days [range, 2–20 days]; P p .03 ). In fatal cases, the median
time to death from the onset of the illness was 10 days (range,
4–24 days). For the 19 children who developed diphtheritic
cardiomyopathy after admission, the examination findings that Figure 3. An 8-year-old boy with diphtheria presenting with nasal
predicted its development were similar to those that predicted pseudomembrane and serous nasal discharge on day 5 of illness.

1594 • CID 2004:39 (1 December) • Kneen et al.

Downloaded from https://academic.oup.com/cid/article-abstract/39/11/1591/463517

by guest
on 02 August 2018
 Table 3. Results of laboratory studies performed at the time of hospital admis-
sion for 154 children with diphtheria.

Children with
Survivors fatal cases
Laboratory value (n p 142) (n p 12) P
Creatinine, mg/dL 0.65 (0.28–1.42) 1.35 (0.9–2.0) !.0001
WBC count, ⫻109 cells/L 16.1 (6.4–34) 12 (12–26.3) .02
Platelet count, ⫻109 platelets/L 170 (108–200) 180 (102–290) .6
Creatinine phosphokinase, IU/L 81.5 (23–2196) 292 (99–3934) .0001
Lactate dehydrogenase, IU/L 278 (74–1303) 386 (253–802) .01
Aspartate aminotransferase, IU/L 37 (12–286) 119 (47–370) !.0001
Cardiac troponin T, ng/mL 0 (0–0.07) 0.09 (0–0.2) !.0001

NOTE. Data are median (range).

severe cases developed signs and chest radiographic abnor-
malities consistent with pulmonary edema.
Admission 12-lead electrocardiography findings. Abnor-
malities of heart rate or rhythm occurred in all 32 children
with diphtheritic cardiomyopathy. Initial 12-lead electrocar-
diographic abnormalities associated with a fatal outcome were
complete heart block (P ! .05 ) and ischemic changes (P ! .01),
whereas sinus tachycardia was associated with a good outcome
(P ! .01) (table 5).
Fatal cases. All patients with fatal cases died of diphtheritic
cardiomyopathy. They all developed worsening heart failure
and died during a cardiac arrest. Two children developed per-
sistent ventricular tachycardia 2 and 4 h before death. The
development of complete heart block occurred in 9 cases; all
were fatal. In 4 patients, complete heart block developed sud-
denly and was the first electrocardiographic abnormality, but
in 5, complete heart block followed other electrocardiographic
abnormalities and was an agonal finding. Four patients (3 with
fatal cases) had a temporary cardiac pacemaker inserted (be-
tween days 7 and 11) as part of a prospective trial [10].
Twenty-four-hour electrocardiography results. A total of
125 continuous 24-h electrocardiographic recordings were
made for 79 patients. Seventy children (89%) had abnormalities
detected on their first 24-h electrocardiograph. At the time of
their first 24-h electrocardiographic recording, 13 children had
diphtheritic cardiomyopathy (symptomatic or asymptomatic),
and all had abnormalities detected. Among the remaining 66
children, results of 24-h electrocardiography at admission were
normal for 47 (71%), if sinus tachycardia is accepted as a variant
of normal, and 1 of these children subsequently developed
diphtheritic cardiomyopathy. Nineteen children had abnormal
24-h electrocardiographic results at admission, and 11 of these
subsequently developed diphtheritic cardiomyopathy (OR,
63.3; 95% CI, 6.6–1507.2; P ! .0001). Thus, the presence of an
abnormality other than sinus tachycardia on a 24-h electro-
cardiograph at admission predicted the development of diph-
theritic cardiomyopathy with a sensitivity of 92%, a specificity
of 85%, a PPV of 58%, and an NPV of 98%. Only 7 of these
children were already predicted to develop diphtheritic cardio-
myopathy by clinical predictors. Thus, 24-h electrocardiogra-
phy improved our ability to predict diphtheritic cardiomyop-
athy by 57%.
Serial 24-h electrocardiography was performed for 39 pa-
tients (32 had 2 recordings and 7 had 3 recordings). For 13

Table 4. Predictors for the development of diphtheritic cardiomyopathy after hospital ad-
mission among 141 children with diphtheria.

Children who Children who

developed did not develop
diphtheritic diphtheritic
cardiomyopathy cardiomyopathy
Examination finding (n p 19) (n p 122) OR (95% CI) P
Classified as “severe” case 12 (63) 9 (7) 21.5 (6–81.6) !.0001
Membrane score of 12 13 (68) 8 (7) 30.9 (8.1–126) !.0001
Bull neck 13 (68) 11 (9) 21.9 (6.1–82.3) !.0001
Bleeding tendency 7 (37) 5 (4) 13.7 (3.2–60.6) .0001
Stridor (requiring tracheotomy) 3 (16) 4 (3) 5.5 (0.9–33.4) .05
Clostridium diphtheriae isolated 12 (63) 13 (11) 14.4 (4.3–50.1) !.0001

NOTE. Data are no. (%) of patients. Thirteen children with diphtheritic cardiomyopathy at presentation are
excluded.

Diphtheritic Cardiomyopathy in Children • CID 2004:39 (1 December) • 1595

Downloaded from https://academic.oup.com/cid/article-abstract/39/11/1591/463517

by guest
on 02 August 2018
 patients, diphtheritic cardiomyopathy either was present at the specificity of 99%, a PPV of 67%, and an NPV of 99%. How-
time of admission or developed after admission. All had more ever, all of these patients were identified by the clinical pre-
significant findings on their second recording. In 2 children, a dictors of fatal outcome.
third 24-h electrocardiograph was recorded after symptoms and Microbiology. Thirty-four patients (22%) had C. diphth-
signs of diphtheritic cardiomyopathy had resolved (days 30 and eriae isolated from their nasopharynx. In all 34 cases, C. diphth-
37 of the illness), and abnormalities were still present on these eriae was isolated from nasal swabs, but only 13 patients (38%)
recordings. had either nasal discharge or pseudomembrane present. C.
Ischemic findings (ST depression or elevation) on the ad- diphtheriae was more likely to be isolated from patients with
mission 24-h electrocardiograph were more frequent in those more extensive pseudomembrane (20 of 34 with positive cul-
who died (8 of 8 patients with fatal cases vs. 5 of 71 of those ture results had pseudomembrane score of 12 vs. 11 of 120
with nonfatal cases; P ! .0001). The median electrocardio- with negative results of culture; P ! .001 ). All of the isolates
graphic abnormality score for admission electrocardiographs were susceptible to penicillin, ampicillin, ceftriaxone, and ri-
was higher for those who died (3.5 [range, 2–8] vs. 0 [range, fampicin. Although there are no clinically relevant break points
0–6]; P ! .0001). for C. diphtheriae, 9 of the isolates showed reduced suscepti-
Laboratory investigation results. Levels of cardiac en- bility or resistance to ⭓1 of the antimicrobials, on the basis of
zymes (creatinine phosphokinase, lactate dehydrogenase, and a high MIC and no zone of inhibition in the disk susceptibility
AST) were higher at the time of hospital admission for patients test. Three isolates were resistant to tetracycline (MIC, 18 mg/
who died (table 3). Of the cardiac enzymes, an elevated AST L); 1 was resistant to trimethoprim (MIC, 8 mg/L); 1 was
level (147 IU/L) was the best predictor of fatal outcome, with resistant to erythromycin (MIC, 116 mg/L) and azithromycin
a sensitivity of 100%, a specificity of 66%, a PPV of 20%, and (MIC, 116 mg/L); 3 were resistant to erythromycin (MIC, 116
an NPV of 100%. mg/L), azithromycin (MIC, 116 mg/L), and tetracycline (MIC,
cTnT results. cTnT levels were measured at the time of 18 mg/L); and 1 was resistant to erythromycin (MIC, 116 mg/
hospital admission for 90 children (7 with fatal cases). Median L), azithromycin (MIC, 116 mg/L), tetracycline (MIC, 18 mg/
cTnT levels were higher at admission in those who died (0.09 L), and chloramphenicol (MIC, 14 mg/L). For 7 (21%) of the
ng/mL [range, 0–0.2 ng/mL] vs. 0 ng/mL [range, 0–0.07 ng/ 34 children with a positive culture result, the isolate was re-
mL]; OR, 160; 95% CI, 11.7–5101.5; P ! .0001). cTnT levels sistant to erythromycin. All of these children were treated with
were elevated (10.01 mg/L) at hospital admission in 6 of the 7 penicillin.
patients with fatal cases in whom the levels were measured. Follow-up. Seventy-nine diphtheria survivors (56%) re-
Therefore, the presence of cTnT in serum samples at hospital turned for a follow-up visit 1 month after discharge (15 were
admission predicted fatal outcome, with a sensitivity of 86%, survivors of diphtheritic cardiomyopathy). All had normal 12-
lead electrocardiographic results. Only 1 child appeared unwell.
He was a survivor of diphtheritic cardiomyopathy, and ex-
Table 5. Initial 12-lead electrocardiographic findings for 32
children with diphtheritic cardiomyopathy.
amination findings revealed sinus tachycardia (rate, 120 beats/
min) and cranial nerve palsies. He had no other signs of cardiac
No. (%) of patients dysfunction, and the sinus tachycardia was assumed to be due
Children with to autonomic neuropathy. Throat and nose swab culture results
Electrocardiographic fatal cases Survivors were negative in all but 1 child—a sibling of the child with a
abnormality (n p 12) (n p 20) P cleft palate [4].
Heart block
First degree 1 (8) 1 (5) .99 DISCUSSION
Second degree 0 (0) 1 (5) .99
Third degree 4 (33) 0 .014 Most descriptions of patients with diphtheria were published
Sinus tachycardia 3 (25) 18 (90) .0003 160 years ago, and their findings concentrated on 12-lead elec-
Sinus bradycardia 0 2 (10) .51 trocardiographic abnormalities. Many of these reports con-
Sinus arrhythmia 0 3 (15) .27 cluded that the development of significant conduction defects,
QTc of 10.44 s 2 (17) 6 (30) .68 such as bundle branch block and complete heart block, had a
VE and/or SVE 4 (33) 8 (40) .99
poor prognosis [6–9, 18]. More recently, 24-h electrocardio-
Bundle branch block 4 (33) 3 (15) .38
graphic monitoring was performed for 15 children with severe
Ischemic changesa 10 (83) 3 (15) .0002
diphtheria from our center and was found to be useful in
NOTE. Data are no. (%) of patients. Some children had 11 abnormal finding predicting fatal outcome [5]. After the recent diphtheria epi-
on electrocardiography. QTc, QT interval corrected for heart rate; SVE, su-
praventricular ectopic; VE, ventricular ectopic. demic in Russia, a study of diphtheritic cardiomyopathy in 122
a
ST depression or T wave inversion. children (9 with fatal cases) found that electrocardiographic

1596 • CID 2004:39 (1 December) • Kneen et al.

Downloaded from https://academic.oup.com/cid/article-abstract/39/11/1591/463517

by guest
on 02 August 2018
 abnormalities with ischemic changes resembling those seen in
adults with an acute myocardial infarction, together with a high
myoglobin level (12000 ng/mL), were reliable markers for fatal
outcome [19]. Other previously identified markers of a poor
outcome in diphtheritic cardiomyopathy include extensive
pseudomembrane in the oropharynx, extensive respiratory tract
infection with subcutaneous edema, leukocytosis (leukocyte
count, 125 ⫻ 10 9 leukocytes/L), and elevated AST level (180
IU/L) [20, 21]. Although previous studies have identified in-
dicators of a fatal outcome, it is not known whether we can
predict the development of diphtheritic cardiomyopathy and
thus focus attention on patients with a potentially treatable
complication, nor is it clear whether newer markers of cardiac
damage, such as presence of cTnT, have anything useful to add.
We therefore conducted a study to identify clinical, electrocar-
diographic, and laboratory features associated with the devel-
opment of cardiomyopathy and fatal outcome.
Overall, 21% of our patients had diphtheritic cardiomyop-
athy, with most (60%) developing signs after admission to the
hospital. Clinical features at presentation provided useful in-
dicators of how patients will fare. The combination of the pres-
ence of a bull neck and a pseudomembrane score of 12 was
the best predictor that patients would develop diphtheritic car-
diomyopathy, whereas the combination of diphtheritic cardio-
myopathy at hospital admission and a pseudomembrane score
of 12 was the best predictor of a fatal outcome.
Ischemic changes on both 12-lead and 24-h electrocardiog-
raphy were more common in patients who died. In addition,
the number of different abnormalities on the 24-h electrocar-
diograph obtained at hospital admission was a good predictor
of fatal cases. The admission 24-h electrocardiograph was also
useful in predicting the development of symptomatic diph-
theritic cardiomyopathy after hospital admission, identifying
190% of such patients.
As others have shown previously, we found that severe con-
duction defects, including complete heart block and bundle
branch block, were associated with a poor outcome [18, 22,
23]. Tachyarrhythmias are reported less often in patients with
diphtheritic cardiomyopathy [24], but we saw sinus arrhythmia,
atrial flutter, prolonged ventricular tachycardia, and several un-
sustained episodes of supraventricular tachycardia. Although
there are rare reports of the development of progressive con-
duction defects years after diphtheria [25, 26], we found no
evidence of persistent cardiac dysfunction in our patients at
follow-up, which included 15 of the 20 survivors of diphtheritic
cardiomyopathy.
cTnT is a thin-filament contractile protein present in high
concentrations in the myocardium and not in other tissues. It
is released rapidly after myocardial injury in direct proportion
to the extent of injury and is undetectable in healthy children
[27]. Preliminary data showed that cTnT levels were elevated
Diphtheritic Cardiomyopathy in Children • CID 2004:39 (1 December) • 1597
Downloaded from https://academic.oup.com/cid/article-abstract/39/11/1591/463517
by guest
on 02 August 2018
in patients with severe diphtheritic cardiomyopathy [10].
Therefore, we wondered whether it might be a useful early
predictor of the development of diphtheritic cardiomyopathy,
but it did not add to the clinical assessment. Other admission
laboratory investigations useful in predicting fatal outcome in-
cluded elevated serum creatinine (for age and sex) and AST
(147 IU/L) levels—both had a high sensitivity and NPV.
In summary, a high risk of developing cardiomyopathy or
of dying can be predicted in children with diphtheria from
both clinical and laboratory investigations and electrocardio-
graphic findings. A simple score to document the amount of
pseudomembrane visible at the time of admission, in combi-
nation with the presence of a bull neck, was the best predictor
of the development of diphtheritic cardiomyopathy. The same
pseudomembrane score in combination with presence of diph-
theritic cardiomyopathy at presentation was the best predictor
of a fatal outcome. Where available, 24-h electrocardiographic
monitoring may help predict the development of symptomatic
diphtheritic cardiomyopathy in some patients, and estimation
of cTnT may help detect subclinical disease. Overall, however,
the simple clinical predictors were as good as the newer cardiac
diagnostic tools. When the availability of cardiac investigations
is limited, these simple clinical predictors will enable doctors
to target resources to the patients that need them most.
Acknowledgments
We thank the directors and staff of the Centre for Tropical Diseases and
the Wellcome Trust Clinical Research Unit, in particular Tran Tinh Hien,
Jeremy Farrar, Delia Bethell, Tran Thi Nhu Thuy, John Wain, Tran Thi My
Van, and Tran Buu Long. We also thank Annie Siemieniuk for biochemical
investigations and Becky Sumner for reporting 24-h electrocardiographs.
Financial support. Wellcome Trust of Great Britain.
Potential conflicts of interest. All authors: No conflicts.
References
1. White NJ, Hien TT. Diphtheria. In: Cook GC, ed. Manson’s tropical
diseases. 20th ed. London: Saunders, 1996:931–5.
2. Rakhmanova AGLJ, Groundstroem K, Valova E, Nosikova E, Tanasi-
jchuk T, Saikku J. Diphtheria outbreak in St. Petersburg: clinical char-
acteristics of 1860 adult patients. Scand J Infect Dis 1996; 28:37–40.
3. Wesley AG, Pather M, Chrystal V. The haemorrhagic diathesis in diph-
theria with special reference to disseminated intravascular coagulation.
Ann Trop Paediatr 1981; 1:51–6.
4. Kneen R, Giao P, Solomon T, et al. Penicillin versus erythromycin in
the treatment of diphtheria. Clin Infect Dis 1998; 27:845–50.
5. Bethell D, Dung N, Loan H, et al. Prognostic value of electrocardio-
graphic monitoring in severe diphtheria. Clin Infect Dis 1995; 20:
1259–65.
6. Smith S. Heart rhythm in diphtheria. JAMA 1922; 77:765–71.
7. Stecher R. Electrocardiographic changes in diphtheria; complete au-
riculoventricular dissociation. Am Heart J 1929; 4:545–58.
8. Hoyne A, Welford N. Diphtheritic myocarditis, a review of 496 cases.
J Pediatr 1934; 5:642–53.
9. Burkhardt E, Eggleston C, Smith L. Electrocardiographic changes and
peripheral nerve palsies in toxic diphtheria. Am J Med Sci 1938; 195:
301–13.
 10. Dung N, Kneen R, Kiem N, et al. Treatment of severe diphtheritic
myocarditis by temporary insertion of a cardiac pacemaker. Clin Infect
Dis 2002; 35:1425–9.
11. Efstratiou A. Diphtheria: manual for the laboratory diagnosis of diph-
theria. Copenhagen: World Health Organization, 1994.
12. Maple PAC, Efstratiou A, Tseneva G, et al. The in-vitro susceptibilities
of toxigenic strains of Corynebacterium diphtheriae isolated in north-
western Russia and surrounding areas to ten antibiotics. J Antimicrob
Chemother 1994; 34:1037–40.
13. Rainbow S, Tickner T. Development of an enhanced chemilumines-
cence assay for troponin-T using Enzymun-test reagents. Clin Chem
1993; 39:1258.
14. Kremer LC, Bastiaansen BA, Offringa M, et al. Troponin T in the first
24 hours after the administration of chemotherapy and the detection
of myocardial damage in children. Eur J Cancer 2002; 38:686–9.
15. Park MK, Guntheroth WG. How to read pediatric ECGS. 3rd ed. St.
Louis: Mosby Year Book, 1992.
16. Begg N. Manual for the management and control of diphtheria in the
European region. Copenhagen: World Health Organization, 1994.
17. Altman DG. Practical statistics for medical research. Padstow, Cornwall,
UK: TJ Press, 1993.
18. Begg ND. Diphtheritic myocarditis: an electrocardiographic study. Lan-
cet 1937; 232:857–60.
1598 • CID 2004:39 (1 December) • Kneen et al.
Downloaded from https://academic.oup.com/cid/article-abstract/39/11/1591/463517
by guest
on 02 August 2018
19. Loukoushkina EF, Bobko PV, Kolbasova EV, et al. The clinical picture
and diagnosis of diphtheritic carditis in children. Eur J Pediatr 1998;
157:528–33.
20. Lumio JT GK, Melnick OB, Huhtala H, Rakhmanova AG. Electrocar-
diographic abnormalities in patients with diphtheria: a prospective
study. Am J Med 2004; 116:78–83.
21. Havaldar PV, Sankpal MN, Doddannavar RP. Diphtheritic myocarditis:
clinical and laboratory parameters of prognosis and fatal outcome. Ann
Trop Paediatr 2000; 20:209–15.
22. Morgan BC. Cardiac complications of diphtheria. Pediatrics 1963; 32:
549–57.
23. Ledbetter MK, Benson Cannon A, Fonesca Costa A. The electrocar-
diogram in diphtheritic myocarditis. Am Heart J 1964; 68:599–611.
24. Neubauer C. Auricular fibrillation and auricular flutter in diphtheria.
Br Heart J 1945; 7:59–62.
25. Butler S, Levine SA. Diphtheria as a cause of late heart-block. Am
Heart J 1930; 5:592–8.
26. Sayers EG. Diphtheritic myocarditis with permanent heart damage.
Ann Intern Med 1958; 48:146–57.
27. Lipshultz SE, Rifai N, Sallan SE, et al. Predictive value of cardiac tro-
ponin T in pediatric patients at risk for myocardial injury. Circulation
1997; 96:2641–8.