Clinical significance of elevated serum ldh|
upon admission in internal medicine.
Authors : F. De Saedeleer, C.-P.R. Yeung, J. Janssens, M. Philippe, C. Beguin, B.
Département de médecine interne, biologie et informatique médicale, Cliniques Universitaires
St-Luc, Université Catholique de Louvain, Avenue Hippocrate 10, 1200 Bruxelles.
Running Titel : Clinical significance of elevated serum ldh upon admission in internal
Address for correspondance : Prof. B. Boland, Service de Médecine Interne, Cliniques
Universitaires ST-Luc, Avenue Hippocrate 10, B-1200 Bruxelles. Boland@mint.ucl.ac.be tél.
++32 2 764 12 57
Serum lactate dehydrogenase (LDH) are routinely measured upon hospital
admission in internal medicine (IM). Their elevation (>340 UI/L) is common : a
recent study (1) in our unit has shown an elevation in more than an half (56 %) of the
incoming patients in IM.
Except for some evident diagnoses, the signignificance of their elevation may
face the attending internist with uncertainties. The clinician has to choice : neglecting
this abnormality or propose to the patients more investigations fearing some severe
desease as neoplasia, lympho-proliferative disorders or unrecognized thrombo-
They are no many study over the rate of these abnomalities and their clinical
significance : one study (2) showed in 1982 an elevation ( in this case > 250 UI/L,
normal values : 75 – 225) in only 100 of 2175 patients (5%), in whom 47 cancers, 19
hepatic lesions, 7 cardiac cases (from whom 5 cardiac livers), 6 cases of hemolyses
and 3 polymyosites.
This study aim to describe the prevalence of an elevation of LDH in the IM
incoming patient and the clinical significance of this abnormality, helping the clinician
making is opinion and taking the good decisions.
This is a prospective and descriptive study lasting for 6 months (november 97
to april 98) in a IM unit. Every patient was eligible and the seric level of LDH was
mesured on the first blood analyse (admission) when arriving in the unit (in some
cases, the patient was transfered from a other unit). When the seric LDH value was
greater than our laboratory normal values ( XX-340 UI/L) the patient entered in the
At this moment the resident, with the help of a student, collected clinical and
biological data. They had also to establish their own hypothesis for this abnormality
and program several basic investigations as chest radiography, electrocardiogram and
The hypothesis had to be choiced in 8 categories : see definitions later. In 99
of the 103 patients, one category was precised by the resident.
Later in the hospitalisation (control day, cD), between the third and the fifht
day, a biological control, more extensive, was realised.
After the discharge, the medical data were reviewed and the etiology,
according to our objective criteria was determined, when possible. Other data, as the
reason for the hospitalisation and the principal diagnose at the discharge were also
Objective criteria were defined to categorise the origine of the LDH elevation.
When it was not possible the origin was classified as « unknown ». When there was
several etiologies, we defined a primary and a secondary etiology, e.g. in the case of
hepatic metastases of a cancer, the cancer was the primary etiology while the liver the
The etiologies were defined in the following way :
* Hemolysis (red blood cells or RBC): haptoglobin < 50 mgr/dl and total bilirubin
>1 mgr/dl or the presence of an obvious clinical etiology as metallic prosthetic
* Heart : abnormal troponin (>0,06 mgr/dl) or an electrocardiogram showing a
* Liver : level hepatic transaminases three times the superior normal value and a
bilirubin >1 mgr/dl, or the presence of an obvious hepatic injury/desease.
* Muscles : fall context or CK three time the superior normal value.
* Cancer : cancer known of confirmed by appropriate clinical investigations.
* Deep vein thrombosis – pulmonary embolism (DVT-PE) : High probability
* Miscellaneous : oesophagal ulcus, pneumonia, pleuresia, vertebral fracture,
pancreatitis, myelo-proliferative disorder, deep vein thrombosis, bed-sore and
These are only descriptive. Because of the biological caracter of these data,
the median value were more representative for the central tendencies and the normal
distribution was checked after a logarythmic transformation.
Patients caracteristics :
Table 1. Caracteristics of the patients.
During these six month 103 of these 319 consecutive patients were included,
according to a abnormal value of the LDH (>340 UI/L). The median value of these
103 patient was 447 UI/L. This elevation was moderate (340 – 400 UI/L) in 30 (29
%) patients and marqued (>400 UI/L) in the 73 (71 %) others.
The median age was 72 and the sex ratio (% female) 0,61 . Approximately 60
% of the reasons for hospitalisation were grouped in repiratory (18 %), infectious
(16.5 %), digestive (15 %) and cardio-vascular (12 %) deseases. The table 1 shows
that these caracteristics were similar in case of moderate and marqued LDH elevation.
The principal diagnose at discharge varied in great extend but 26 % of the
cases were related to a respiratory illness, 10 % to a digestive illness, 8 % to a
neoplastic illness and 8 % to a cardio-vascular illness.
Table 2 : median level at the admission and at the control-day ( UI/L).
The levels of the other enzymatical values were analysed. The levels of the
creatine kinase (CK) and transaminases were normal in 75 % of these cases. The
levels of the alcaline phosphatase and gammaglutamyl transférase were also normal in
respectively 95 and 50 % of the cases. The median values of these enzymes at the
admission and at the control day are in the table 2. The tendency goes generaly to the
normalisations of these values.
Etiology of the LDH elevation :
The hypothesis formulated by the resident was in order of frequency an origin
attributed to the liver (23 %), muscles (14 %), cancer (9 %), miscellaneous (9 %),
RBC (8%), DVT-PE (7 %), heart (5 %) and unknown (22 %).
In comparison, the etiology, at the reviewing of the clinical data, according our
objective criteria, was considered to be cancer (18 %), muscles (13 %), miscellaneous
(11 %), RBC (7 %), DVT-PE (6%), liver (3 %) or heart (1%). In 42 % of these cases
the etiology was not precisable, thus classified as unknown. As described elsewhere
(3) a non-substitutive corticoïd treatment was considered as a potential origin for an
elevation of the LDH level, but of secondary importance. This occured in 15 % of the
Table 3 : Initial hypothesis according to etiology.
The table 3 approachs the sensibility of the clinical suspicion, since he shows
the proportion of initial hypothesis that were exact. Globally, 66 % of the etiologies
were correctly suspected. This was the case for the majority of the origins attributed
to the muscles (67 %), the RBC (71 %), the DVT-PE (83 %) and the few cases of
liver and cardiac origins. In the other way the cancer origin was suspected only in 44
% of cancer cases.
Table 4 : Etiology according to initial hypothesis.
The table 4 shows, for each category of hypothesis, the proportion of exact or
erroneous suspicion. He approach thus the positive predictive value of the clinical
suspicion. Globally, the hypothesis was confirmed in half the cases (49 %), but it was
confirmed in two third of the cases (66 %) when a hypothesis could be proposed. The
initial hypothesis was correct in the majority of the suspicion for the cases attributed to
cancer (73 %), muscles (57 %), RBC (63 %), DVT-PE (71 %) whereas the origin
attributed to the heart (20 %) and the liver (13 %) were seldom confirmed.
Within the 30 patients whose LDH’s elevation was moderated, the main
hypothesis were also an origin from the liver (27 %) and the muscles (13 %). Also the
proportion of the « unknown » was nearly a quart (23 %) of the cases. Also the
origins from the liver and cancers were respectively over and underrated. There was
no specific etiology, according our objective criteria, in 17 (57 %) of these 30 cases.
Table 5 : level of LDH at admission according to etiology.
The table 5 shows, for each etiology, the LDH’s median value at the admission
and the proportion of moderated versus marqued elevations. For exemple, in the
cancer cases, the LDH’s elevation was moderated in 26 % and marqued in 74 % of the
Evolution of the abnomalities.
On the control day, the value of the LDH remained elevated in 77 % of the
cases ( moderately in 22 % and marquedly in 55 %). The global median value has
lessen to 417 UI/L versus 447 UI/L, a difference of 7 %.
Table 6 : Evolution of the LDH’s level according to the etiology
The evolution was different according to the etiology, as shown in the table 6.
The LDH’s level was frequently normalised in three categories of origin : muscles (23
%), miscellaneous (27 %), DVT-PE ( 17 %) and in a much lesser extend in the cases
of cancer ( only 10 %). When the origin was the RBC, the LDH’s level remainded
abnormal. The origin was to seldom the liver or the heart to permit an analysis.
As expected, a majoration of LDH value is frequent. In this study we
measured it globally at 32 %, with 9.4 % being of moderate elevation. This
abnomality is thus less frequently observed than in the study of 1997 (1) but much
more frequently than in the study of 1982 (2), although the normal values were more
severe in these last study.
The correlation between the hypothesis and the etiology is rather good when
the diagnose is clear (e.g. traumatisms), also, when the suspected origin is cancer it
has many chances to be confirmed, maybe because there are many other clinical signs
as anorexy, weight loss,..
In our study, when the hypothesis is an origin from the liver, we see that it is
frequently a cancer origin. Globally, the cancer origin is underrated as only 44 % of
the cases were suspected. Maybe in the presence of other biological abnomalities (in
whom the "liver" enzymes) the attention is drawn to the liver.
In many cases the clinician has no hypothesis and we see that he is under the
reality because when the data are reviewed, according our objective criteria, as many
as 42 % of the cases still don’t have a clear origin. Some pathologies may pass
unnoticed because of the context, all the attention of the clinician being monopolysed
by some more vital considerations.
It seems to be obvious that the level of this enzyme is highly variable in the
time and that, consecutively, the level of the dosage mainly depends on the moment at
witch the prelevement is done, in relation to the moment on witch the injury take
place. It also seems obvious that the amplitude of the elevation is also important,
determining the duration of the period in witch the level will be abnormal, although it
may greatly varies : this is the explanation why some cases of muscular pathology, of
traumatic origin do not normalise in a few days. A later control would had be more
Furthermore, the muscular pathology is generally underrated in the reasons for
hospitalisation and the final diagnoses : this pathology is then an epiphenomenon,
taking place in the context of a fall, for exemple.
In the intermediate level, as execpted, it was more difficult to draw
conclusions. But in these patients the proportion of cancer (16.7 %) and DVT-PE is
not negligible too.
It is thus effectively a frequent abnomality, often relatively lonely and difficult
to interprete, even at the end of the observation. Even the prevalence of this
abnomality is highly variable from an observation to an other. The consequence of this
is the difficulties to establish a cut-off, the tendency to raise it at a level at witch the
origin is clearer and thus loosing some sensibility. Although, it would be a pity to
neglect his informative value as it is commonly realised.
The level is highly variable, according to the patient, his age and his pathology,
but also the moment in the history of his illness. It is thus important to integrate this
biological information in the clinical context.
In many cases, the elevation goes to a normalisation but this normalisation is
less frequent in the cases of thrombo-emolism, hémolysis and cancer.
Thus if we don’t observe a normalisation it would be rentable to propose some
basic, not very traumatic investigations as a pulmonary scintigraphy and an abdomen
ultrasonoscopy because it may revelate an underlying condition that require a rapid
Serum lactate dehydrogenase (LDH) are routinely measured upon hospital
admission in internal medicine (IM).
Their elevation (>340 IU/L) is common, has multiple possible severe etiologies, and
may face the attending internist with uncertainties.
Serum LDH were measured upon admission (day 1) in 319 consecutive patients of
our IM unit during six months in 1998. The 103 cases (32%) with elevated LDH were further
evaluated (mean age 67 yrs ; women 61%). On day 2, the IM resident was asked about her/his
explanation for LDH elevation. On the control day , blood tests ( haptoglobin, troponin, CPK,
SGOT, SGPT, bilirubin) and liver ultrasound were performed. After hospital discharge, the
medical record was reviewed (e.g. medical history, final diagnoses) to search for the LDH
etiology according to 7 pre-defined diagnostic categories : red blood cells (RBC), heart,
muscle, liver, thrombo-embolism (DVT-PE), cancer, and unknown.
On day 1, median LDH was 447 IU/L in the 103 patients. On day 2, IM residents
suggested that LDH elevation was due to a liver (23%), muscle (14%), cancer (11%), RBC
(8%), DVT-EP (7%) or heart (5%) disease, or had no hypothesis (31%). On day 3, mean LDH
had decreased by 12% but remained elevated in 80 (78%) patients. After discharge, 56 cases
(54%) were still unexplained, while a definite cause was found in 47 cases: cancer (17%),
muscle (12%), RBC (7%), DVT-PE (6%), liver (3%) and heart (1%). The resident hypothesis
on day 2 had a 64% global true positive rate (i.e. sensitivity) as they correctly identified 30 of
the 47 final diagnoses. They had missed 1 of 6 DVT-PE but 10 of 18 cancer diagnoses. LDH
decrease was smaller in the cancer cases (4 vs 15%) on day 3.
Cancer is a main cause of unexplained LDH elevation that should be
searched for particularly if high LDH is confirmed. The significance of LDH of
unknown origin deserves further studies.
1. B. Boland, S. Thomas, F. Goffin, C. Beguin et M. Lambert. Pertinence du dosage
systématique de la lipase sérique lors de l’hospitalisation en médecine interne.
Louvain Méd. 1997 ;116 : 301-311.
2. Ph. Arlet, Y. Botreau, M. Duffaut et Y. Le Tallec. Intérêt et signification
diagnostique de la lacticodéshydrogénase plasmatique en médecine interne. Sem.
Hôp. Paris 24 juin 1982 ; 58 (25) :1343 – 8
3. Selvais PL, Maiter DM, Schoemans H, Seidemann I, Lambert MJ. Serum lactate
dehydrogenase activity increases in both endogenous and exogenous
hypercorticisms. Acta Clin Belg 2000 Jul-Aug;55(4):205-8
Création le 1 avril 2003.
Dernière mise à jour le dimanche 4 juillet 2004.
issu de ldh17avril.doc - 22/04/2001