Prognosis of patients presenting extreme acidosis (pH < 7) on admission to
Intensive Care Unit

Jérôme Allyn MD, David Vandroux MD, Julien Jabot MD, Caroline
Brulliard MD, Richard Galliot MD, Xavier Tabatchnik MD, Patrice Combe
MD, Olivier Martinet MD, Nicolas Allou MD

PII: S0883-9441(15)00506-7
DOI: doi: 10.1016/j.jcrc.2015.09.025
Reference: YJCRC 51962

To appear in: Journal of Critical Care

Please cite this article as: Allyn Jérôme, Vandroux David, Jabot Julien, Brulliard Caro-
line, Galliot Richard, Tabatchnik Xavier, Combe Patrice, Martinet Olivier, Allou Nicolas,
Prognosis of patients presenting extreme acidosis (pH < 7) on admission to Intensive Care
Unit, Journal of Critical Care (2015), doi: 10.1016/j.jcrc.2015.09.025

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Prognosis of patients presenting extreme acidosis (pH <7) on admission to Intensive Care Unit.

a,*
Jérôme Allyn, MD , David Vandroux, MD a, Julien Jabot, MD a, Caroline Brulliard, MD a, Richard Galliot,

MD a, Xavier Tabatchnik, MD a, Patrice Combe, MD b, Olivier Martinet, MD a, Nicolas Allou, MD a

T
a
Intensive care unit, CHU Félix Guyon, La Réunion, Bellepierre 97405 Saint-Denis cedex, France.

P
b
Biology laboratory, CHU Félix Guyon, La Réunion, Bellepierre 97405 Saint-Denis cedex, France.

RI
SC
* Corresponding author:

Dr Jérôme Allyn, Réanimation Polyvalente, CHU La Réunion Félix Guyon, Bellepierre 97405 Saint-Denis

NU
cedex, France.
MA
Phone number: 00262 262905690

Fax number: 00262 262906693

ED

E-mail: allyn.jer@gmail.com
PT
CE
AC

1
 ACCEPTED MANUSCRIPT

Abstract

Purpose: To determine prognosis of patients presenting extreme acidosis (pH <7) on admission to Intensive

Care Unit (ICU); and to identify mortality risk factors.

T
Materials and methods: We retrospectively analyzed all patients who presented with extreme acidosis within

P
24 hours of admission to a polyvalent ICU in a University hospital between January 2011 and July 2013.

RI
Multivariate analysis and survival analysis were used.

SC
Results: Among the 2,156 patients admitted, 77 patients (3.6%) presented extreme acidosis. Thirty (39%)

patients suffered cardiac arrest before admission. While the mortality rate predicted by severity score was 93.6%,

NU
death occurred in 52 cases (67.5%), in a median delay of 13 (5-27) hours. Mortality rate was depended on reason

for admission, varying between 22% for cases linked to diabetes mellitus, and 100% for cases of mesenteric
MA
infarction (P=0.002), cardiac arrest before admission (P<0.001), type of lactic acidosis (P=0.007), high SAPS II

(P=0.008), and low serum creatinine (P=0.012).

ED

Conclusions: Patients with extreme acidosis on admission to ICU, have a less severe than expected prognosis.

While mortality is almost 100% in cases of cardiac arrest before admission, mortality is much lower in the
PT

absence of cardiac arrest before admission, which justifies aggressive ICU therapies.
CE

Keywords: Acidosis-Lactate-Prognosis
AC

2
 ACCEPTED MANUSCRIPT

1. Introduction

Acidosis is a potentially life-threatening abnormality of the acid-base balance that has effects on the

cardiovascular, pulmonary and immune functions [1,2]. The causes can be many and varied, with each requiring

specific treatments. Otherwise, acidosis itself may require symptomatic treatment in severe cases [3-5].

P T
The existing literature on acidosis in intensive care unit (ICU) mainly consists of analyses of specific causes of

RI
metabolic acidosis, with little on management of severe cases; there is no study comparing different causes [6–

SC
9]. As regards to shock with severe lactic acidosis with pH <7.2, observed mortality rate is about 50%, while

several study have reported universal mortality in case of shock with lactic acidosis under pH 7.0 [10–13]. Only

NU
one study examined the causes and prognosis of patients with extreme acidosis in intensive care, but this study

included only 28 patients, and pH criterion, at 6.8, was very extreme [14].
MA
The prognosis of patients with extreme acidosis on admission to ICU is unknown, and it is likely that the cause

of acidosis and management in ICU is related to mortality. The objective of this study was to describe causes and
ED

prognosis of patients with extreme acidosis (pH <7) admitted to ICU, and to identify mortality risk factors.
PT

2. Materials and methods

CE

2.1. Study population

We conducted a retrospective observational study, in a 23-bed adult medical-surgical ICU in a French University
AC

hospital between January 2011 and July 2013. The study was approved by the Ethics Committee of Reunion

Island University Hospital (reference R14007), and no consent was need. We retrospectively analyzed all

patients who presented with acidosis with an arterial pH <7 during the first 24 hours after admission to ICU.

Patients with arterial pH <7 were identified via arterial blood gas results obtained from the hospital biochemistry

department. Our ICU does not currently have a management protocol for such patients.

2.2. Data collection and processing

Demographic data, test results, therapeutics used and patient progress were analyzed. The primary outcome was

the mortality rate during hospitalization in ICU. Therapeutics used included renal replacement therapy,

vasoactive drugs and intravenous administration of sodium bicarbonate.

3
 ACCEPTED MANUSCRIPT

Blood gas analyses were performed on a Cobas b 221 system (Roche Diagnostics GmbH, Mannheim, Germany)

with the integrated AutoQC drawer option and electrodes for PO 2, PCO2, pH, a reference electrode and an ISE

electrode measuring Na+, K+ and Cl-. Specimens were analyzed using our Standard Operating Procedure which

included preanalytical protocols (such as safety, temperature and transportation time) by trained authorized

T
laboratory technicians, with automated internal quality control and direct connection to our laboratory software

P
24 hours a day. External quality controls were performed at least four times a year with samples.

RI
2.3. Definitions

SC
Extreme acidosis was arbitrary defined by an arterial pH of <7.00. Acidosis was arbitrary considered as

NU
corrected if arterial pH was greater than 7.30.

Cases of acidosis were classified as metabolic (PaCO2 <40 mmHg and HCO3- <24 mmol.l-1), respiratory (PaCO2
MA
>40 mmHg and HCO3- >24 mmol.l-1) and mixed (PaCO2 >40 mmHg and HCO3- <24 mmol.l-1) [15].

Causes of acidosis were determined by analysis of patients’ medical charts. Criterion suggesting metformin
ED

associated lactic acidosis (MALA) was the presence of metabolic lactic acidosis with arterial lactate higher than

5 mmol.L-1, no other cause of elevated lactate, and a metformin accumulation assessed by plasma levels, if
PT

available [6].
CE

Cases of acidosis with arterial lactate > 4 mEq.l-1 were classified into two categories (Type-A and Type-B),

based on the presence (Type-A) or not (Type-B) of an evident impaired tissue oxygenation [16].
AC

2.4. Statistical analysis

Results are expressed as median (first and third quartile) and numbers (proportions) as appropriate. Ninety-five

percent confidence intervals are provided. Continuous variables were compared using the Mann-Whitney U test

and categorical variables using the Chi-square test or Fischer’s exact test, as appropriate. A P value less than

0.05 was considered significant.

The risk factors found to be predictive of mortality in the bivariate analysis with P<0.05 were entered into a

multivariate logistic regression analysis with backward selection with P=0.05. Model calibration was assessed

using the Hosmer-Lemeshow test, and the Nagelkerke and Cox/Snell R squares were calculated. Analysis was

performed using SPSS version 15 (SPSS Inc, Chicago, Ill, USA).

4
 ACCEPTED MANUSCRIPT

ICU survival rates were analyzed and comparisons were made between patients having suffered cardiac arrest

before admission to ICU and those that had not, using the Kaplan-Meier estimator and the LogRank test.

T
3. Results

P
RI
During the 30-month study period, 2,156 patients were admitted to ICU. Among them, 77 (3.6%) had extreme

acidosis with an arterial pH <7 in the first 24 hours after admission, forming our study cohort.

SC
3.1. Characteristics of patients on admission to ICU

NU
The clinical and biologic characteristics of the 77 patients are presented in Table 1. The median age was 57 (47-

69) years old and the median simplified acute physiology score (SAPS) II at admission was 82 (69-93).
MA
3.2. Type and cause of acidosis

Median pH was 6.94 (6.86-6.97), PaCO2 was 44 (30-64.5) mmHg, HCO3- was 8.6 (6.1-13.4) mmol.l-1, and
ED

median plasma arterial lactate was 15.75 (9.9-20.45) mmol.l-1.

PT

Thirty patients (39%) had suffered cardiac arrest prior to being admitted to ICU, including 9 cases of hypoxic

cardiac arrest (3 drowning, 1 hanging, 3 pneumonia, 1 cardiogenic pulmonary edema, 1 severe acute asthma), 8
CE

cardiac arrests of cardiac origin (including 3 cases of pulmonary embolism), 4 cases of hemorrhage (2 cases of

gastrointestinal bleeding, 1 postpartum hemorrhage, 1 traumatic splenic injury), 1 case of septic shock, 1 case of
AC

thrombotic microangiopathy, 1 case of thiamine deficiency, 1 case of necrotizing pancreatitis caused by

alcoholism, 1 case of fulminant hepatitis, 1 case of mesenteric infarction and 3 cases of unknown origin.

Among the 47 patients (61%) who did not suffer cardiac arrest before admission to ICU, there were 14 patients

with septic shock, 7 cases of MALA, 7 cases of isolated hypercapnic respiratory failure, 4 cases of mesenteric

infarction, 3 cases of cardiogenic shock, 3 cases of hemorrhagic shock, 2 cases of diabetic ketoacidosis, 1

thiamine deficiency, 1 case of necrotizing pancreatitis caused by alcoholism, 1 case of thrombotic

microangiopathy, 1 case of status epilepticus, 1 hypovolemic shock related to diarrhea, 1 case of unknown shock

in a cirrhotic patient and 1 case of unknown shock in a patient with a bronchopulmonary adenocarcinoma.

5
 ACCEPTED MANUSCRIPT

Among the 12 patients who were treated with metformin, acidosis was attributed solely to MALA in 7 cases

while in 5 cases, other causes were identified (2 cases of septic shock, 1 pulmonary embolism, 1 myocardic

infarction, 1 gastrointestinal bleeding).

Arterial lactate was greater than 4 mEq.l-1 in 66 cases (86%), classified as Type-A in 57 cases and Type-B in 9

P T
cases (7 cases of MALA, 1 case of thiamine deficiency, and 1 case of status epilepticus). Only six patients

RI
(7.8%) had no metabolic cause in their acidosis (Table 1).

SC
3.3. Therapy and follow-up

Seventy four patients (96%) were under mechanical ventilation and 74 (96%) under vasopressors. Extra renal

NU
replacement therapy was used in 40 cases (52%), sodium bicarbonate in 26 cases (34%), and extracorporeal

cardiopulmonary support in 4 cases (5%).

MA
pH could be corrected in 32 cases (25 survivors and in 7 deceased), within a median time delay of 17 (10-22)

hours. With survivors, the median time delay for pH >7.3 was 17 (9.5-22) hours. In the group of patients who
ED

died, the pH was >7.3 in 19 (13.5-25) hours (P=0.48).

3.4. Mortality
PT

Death occurred in 52 cases (67.5%). Only 3 patients survived of the 30 who had suffered cardiac arrest before
CE

ICU admission (mortality rate of 90%, 1 case of B1 deficiency, 1 case of thrombotic microangiopathy and 1 case

of cardiogenic shock). In these all 3 cases, the cardiac arrest had occurred in the presence of a medical team, who
AC

immediately started resuscitation (“no flow” 0 minute in all cases). Two of these three patients were relatively

young (35 and 37 years old). The 27 deaths, among these 30 patients, occurred from refractory shock in 17 cases

and neurological causes in 10 cases. Death occurred in 13 (5-38) hours.

Of the 47 patients who did not suffer cardiac arrest before ICU admission, 25 patients (53%) died in 12 (5-17)

hours. Cause of death was refractory shock in 22 cases and a decision to limit therapeutics in 3 cases (i.e. death

after an ICU stay of 1, 2, and 17 days).

Mortality was statistically associated with the presence of lactic acidosis (P=0.004), and also with the type of

lactic acidosis (P <0.001).

Figure 1 presents the Kaplan Meier analysis: survival was significantly higher in patients who did not suffer

cardiac arrest before admission to ICU (LogRank, P=0.003).

6
 ACCEPTED MANUSCRIPT

Within the overall population of 77 patients, mortality was significantly related to SAPS II (P <0.001), incidence

of cardiac arrest prior to admission (P =0.001). Mortality was also inversely related to serum creatinine level and

prothrombin activity (P =0.01), presence of diabetes mellitus (P =0.01). After exclusion of patients with B-type

lactic acidosis, arterial lactate was found to be related to ICU mortality (P =0.04).

P T
Mortality varied significantly depending on the cause of acidosis: 100% in cases of hypoxic cardiac arrest prior

RI
to admission or mesenteric infarction; 0% in cases of diabetic ketoacidosis, thiamine deficiency or thrombotic

microangiopathy (P =0.002). Mortality rates in cases of septic, cardiogenic and hemorrhagic shock were 66%,

SC
73% and 86%, respectively (Table 2). In the overall population, the mortality rate predicted by SAPS II was

93.6%. Figure 2 presents mortality rates for causes of ICU admission compared with the mortality rate predicted

NU
by SAPS II [17]. Mortality rate of patients with A-Type Lactic acidosis, B-Type lactic acidosis, and patients

without lactic acidosis were 82.4%, 22.2% and 27.3%, respectively; while their predicted mortality were 94.7%,
MA
90.1% and 80.6% respectively.

Variables with a p-value <0.05 were entered into the logistic regression (Table 3): diabetes mellitus, serum
ED

creatinine, prothrombin activity, SAPS II, A-Type lactic acidosis. The “Cardiac arrest prior to admission to ICU”

variable was excluded because of multicollinearity. Three variables were significantly associated with an
PT

increased risk of mortality: presence of a A-Type lactic acidosis, serum creatinine and SAPS II.
CE

4. Discussion
AC

To our knowledge, our study of patients with extreme acidosis is the largest recorded. Paz et al., whose study

was recently published, involved 28 patients with a more severe acidosis criterion of pH <6.80. This study

reported only 9 survivors (mortality rate: 68%) [14]. In agreement with these results, we report an overall

mortality rate of 67.5%, lower than SAPS II could predict (93.6%) [17].

Mortality rate in patients having suffered cardiac arrest prior to ICU admission was very high (90%), with

survival only when cardiac arrest occurred in the presence of a medical team, and in relatively young patients.

Our results are in agreement with Donnino et al., a prospective study of lactate levels following cardiac arrest,

which recorded a mortality rate of 80% in patients where initial lactate was greater than 10 mmol.l-1 [8].

7
 ACCEPTED MANUSCRIPT

Furthermore, survival analysis shows that the deaths occur almost exclusively in the first 2 days in patients who

did not suffer cardiac arrest before admission. This supports a very aggressive initial management in ICU during

the first 48 hours of care.

We highlight the critical role of the cause of ICU admission, and the possible occurrence of cardiac arrest prior

P T
to admission as essential prognostic criteria. In addition, multivariate analysis identified three variables

RI
independently associated with mortality: presence of an A-Type lactic acidosis, creatinine and SAPS II.

Surprisingly, there was an inverse relationship between mortality and serum creatinine, which had a protective

SC
effect against death. To explore this point, we performed a specific statistical analysis. We found that patients

who had suffered cardiac arrest before admission had a lower value of serum creatinine than patients who did not

NU
suffer cardiac arrest (158 (130-228) micromol.l-1), versus 229 (138-406) micromol.l-1, P=0.03). It is likely that

serum creatinine on admission does not accurately reflect renal function. Unfortunately, creatinine clearance
MA
cannot be assessed in such acute pathology. Therefore, we think that interpreting creatinine results is difficult

and should be done with caution.

ED

We report a mortality rate of 29% in cases of MALA. Seidowsky’s retrospective study focused on 42 ICU

patients with MALA [6]. Mean pH was 7.1 and mortality was 33%; pH was related to mortality in bivariate
PT

analysis, but not in multivariate analysis.

CE

Moreover, we found a significant association between mortality and presence and type of lactic acidosis, and

also a significant association between mortality and arterial lactate after exclusion of patients with B-type lactic
AC

acidosis. We did not find evidence of changes in mortality rates with differing PaCO2, an easily modifiable

factor under mechanical ventilation. Pure respiratory acidosis was a very marginal element of the overall

population; the most frequent case was mixed acidosis, probably because of the circulatory effects of acidosis

leading to production of lactate or associated renal failure [18].

We found no significant association between mortality and ICU therapy used or with the delay in acidosis

correction. The mortality difference observed regarding the renal replacement therapy was close to be significant

(P =0.06).

Our study has some limitations including the retrospective nature of the analysis and a low statistical power. Our

retrospective approach did not permit a fine biochemical analysis of acidosis, for example by the Stewart

method, which requires albumin values [19,20]. However, for some authors, this “modern” approach does not

8
 ACCEPTED MANUSCRIPT

change the applied management of patients [21,22]. Our study can be considered as a pragmatic approach to the

problem.

On the methodology of the study, we selected only arterial samples (arterial catheter only), and pH threshold we

established (7.00) was based on studies reporting a very severe mortality in case of shock with lactic acidosis

T
under pH 7.0 [10–13]. We found that with a value of 7.00, prevalence was not exceptional because it almost

P
corresponds to the 4th percentile in our ICU.

RI
Moreover, we present 7 patients with MALA diagnosis based on a retrospective analysis of charts but without

SC
available metformine plasma levels.

5. Conclusions NU
MA
Patients admitted to ICU with extreme acidosis have a bleak prognosis but less than expected in absence of

cardiac arrest before admission. While mortality is almost universal in patients suffering cardiac arrest before
ED

admission, mortality is much lower in the absence of cardiac arrest before admission, which justifies aggressive

ICU therapy.
PT
CE
AC

9
 ACCEPTED MANUSCRIPT

Competing interest: All authors declare that they have no competing interests.

Authors’ contributions:

T
JA designed the study, contributed to acquisition, analysis and interpretation of data; drafted the manuscript and

P
RI
has given final approval of this version; DV contributed to analysis and interpretation of data; drafted the

SC
manuscript and has given final approval of this version; JJ contributed to analysis and interpretation of data,

drafted the manuscript and has given final approval of this version; CB contributed to analysis and interpretation

NU
of data, drafted the manuscript and has given final approval of this version; RG contributed to acquisition,
MA
analysis and interpretation of data; drafted the manuscript and has given final approval of this version; XT

contributed to acquisition, analysis and interpretation of data; drafted the manuscript and has given final
ED

approval of this version; PC contributed to analysis and interpretation of data, revised the manuscript critically
PT

for important intellectual content and has given final approval of this version; OM contributed to analysis and
CE

interpretation of data, revised the manuscript critically for important intellectual content and has given final

approval of this version; NA designed the study, contributed to acquisition, analysis and interpretation of data;
AC

drafted the manuscript and has given final approval of this version. All authors read and approved the final

manuscript.

Acknowledgments: Andrew Hobson provided assistance for the English language review, which was funded by

the Universitary Hospital.

10
 ACCEPTED MANUSCRIPT

References

[1] Mitchell JH, Wildenthal K, Johnson RL. The effects of acid-base disturbances on cardiovascular and

pulmonary function. Kidney Int 1972;1:375–89.

[2] Kellum JA, Song M, Li J. Science review: extracellular acidosis and the immune response: clinical and

P T
physiologic implications. Crit Care 2004;8:331–6.

RI
[3] Kraut JA, Kurtz I. Use of base in the treatment of acute severe organic acidosis by nephrologists and

critical care physicians: results of an online survey. Clin Exp Nephrol 2006;10:111–7.

SC
[4] Mathieu D, Neviere R, Billard V, Fleyfel M, Wattel F. Effects of bicarbonate therapy on hemodynamics

and tissue oxygenation in patients with lactic acidosis: a prospective, controlled clinical study. Crit Care

NU
Med 1991;19:1352–6.

[5] Duhon B, Attridge RL, Franco-Martinez AC, Maxwell PR, Hughes DW. Intravenous sodium bicarbonate
MA
therapy in severely acidotic diabetic ketoacidosis. Ann Pharmacother 2013;47:970–5.

[6] Seidowsky A, Nseir S, Houdret N, Fourrier F. Metformin-associated lactic acidosis: a prognostic and
ED

therapeutic study. Crit Care Med 2009;37:2191–6.

[7] Kaplan LJ, Kellum JA. Initial pH, base deficit, lactate, anion gap, strong ion difference, and strong ion gap
PT

predict outcome from major vascular injury. Crit Care Med 2004;32:1120–4.

[8] Donnino MW, Andersen LW, Giberson T, Gaieski DF, Abella BS, Peberdy MA, et al. Initial lactate and
CE

lactate change in post-cardiac arrest: a multicenter validation study. Crit Care Med 2014;42:1804–11.

[9] Gershengorn HB, Iwashyna TJ, Cooke CR, Scales DC, Kahn JM, Wunsch H. Variation in use of intensive
AC

care for adults with diabetic ketoacidosis*. Crit Care Med 2012;40:2009–15.

[10] Noritomi DT, Soriano FG, Kellum JA, Cappi SB, Biselli PJC, Libório AB, et al. Metabolic acidosis in

patients with severe sepsis and septic shock: a longitudinal quantitative study. Crit Care Med

2009;37:2733–9.

[11] Smith I, Kumar P, Molloy S, Rhodes A, Newman PJ, Grounds RM, et al. Base excess and lactate as

prognostic indicators for patients admitted to intensive care. Intensive Care Med 2001;27:74–83.

[12] Bakker J, Nijsten MW, Jansen TC. Clinical use of lactate monitoring in critically ill patients. Ann

Intensive Care 2013;3:12.

[13] Friesecke S, Abel P, Roser M, Felix SB, Runge S. Outcome of severe lactic acidosis associated with

metformin accumulation. Crit Care 2010;14:R226.

11
 ACCEPTED MANUSCRIPT

[14] Paz Y, Zegerman A, Sorkine P, Matot I. Severe acidosis does not predict fatal outcomes in intensive care

unit patients: a retrospective analysis. J Crit Care 2014;29:210–3.

[15] Guyton, Hall. Précis de physiologie médicale. vol. Deuxieme edition francaise. Piccin. Piccin; 2003.

[16] Cohen RD, Woods HF. Lactic acidosis revisited. Diabetes 1983;32:181–91.

T
[17] Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a

P
European/North American multicenter study. JAMA 1993;270:2957–63.

RI
[18] Kimmoun A, Novy E, Auchet T, Ducrocq N, Levy B. Hemodynamic consequences of severe lactic

SC
acidosis in shock states: from bench to bedside. Crit Care 2015;19:175.

[19] Stewart PA. Modern quantitative acid-base chemistry. Can J Physiol Pharmacol 1983;61:1444–61.

NU
[20] Gunnerson KJ, Saul M, He S, Kellum JA. Lactate versus non-lactate metabolic acidosis: a retrospective

outcome evaluation of critically ill patients. Crit Care 2006;10:R22.

MA
[21] Moviat M, van Haren F, van der Hoeven H. Conventional or physicochemical approach in intensive care

unit patients with metabolic acidosis. Crit Care 2003;7:R41–5.

[22] Carreira F, Anderson RJ. Assessing metabolic acidosis in the intensive care unit: does the method make a
ED

difference? Crit Care Med 2004;32:1227–8.

PT
CE
AC

12
 ACCEPTED MANUSCRIPT

Figure 1 Kaplan-Meier ICU survival curve for patients who suffered cardiac arrest before admission to

ICU, and those who did not (Log Rank, P=0.003).

T
Figure 2 Observed and predicted mortality rates, determined by cause of admission in patients who did

P
not suffer cardiac arrest before ICU admission (n=47).

RI
*including MALA (n=7) and diabetic ketoacidosis (n=2).

SC
NU
MA
ED
PT
CE
AC

13
 ACCEPTED MANUSCRIPT

Table 1 Demographic, clinical and biological characteristics of patients. Therapeutics in ICU.

Total Alive Dead P

(n=77) (n=25) (n=52)

Age (year) 57 (47-69) 59 (43-71) 56.5 (48-69) 0.76

T
Gender, male, n (%) 50 (64.9) 17 (68) 33 (63.5) 0.8

P
Hypertension, n (%) 34 (44.2) 13 (52) 21 (40.4) 0.46

RI
COPD, n (%) 4 (5.2) 2 (8) 3 (5.8) 0.66

Malignancy, n (%) 6 (7.8) 1 (4) 6 (11.5) 0.42

SC
Congestive heart failure, n (%) 16 (20.8) 9 (36) 10 (19.2) 0.16

Alcoholism, n (%) 34 (44.2) 10 (40) 24 (46.2) 0.63

Smoking, n (%) 18 (23.4) 9 (36) 9 (17.3) 0.09

NU
Diabetes mellitus, n (%) 24 (31.2) 13 (52) 11 (21.2) 0.01

 With metformine, n (%) 12 (15.6) 8 (32) 4 (7.7)

MA
 Without metformine, n (%) 12 (15.6) 5 (20) 7 (13.5)

Chronic renal failure, n (%) 8 (10.4) 2 (8) 6 (11.5) 1

Hepatic cirrhosis, n (%) 6 (7.8) 0 6 (11.5) 0.17

ED

SAPS II 82 (68.8-93) 69 (55-79) 88 (78-95.5) < 0.001

Cardiac arrest before ICU admission, n (%) 30 (39) 3 (12) 27 (51.9) 0.001

 No flow (min)£ 0 (0-5) 0 (0-0) 1.5 (0-5) 0.15

PT

 Low flow (min)§ 20 (7.5-30) 3 (2.5-4) 20 (10-31.3) 0.01

pH 6.94 (6.86-6.97) 6.96 (6.89-6.98) 6.92 (6.84-6.96) 0.09

CE

PaCO2 (mm Hg) 44 (30-64.25) 42 (29-55) 46 (32-64.5) 0.57

Type of acidosis, n (%) 0.34*

AC

 Metabolic (PaCO2<40 mmHg and HCO3-<24 31 (40.3) 12 (48) 19 (36.5)

mmol.l-1)

 Respiratory (PaCO2>40 mmHg and HCO3->24 6 (7.8) 4 (16) 2 (3.8)

mmol.l-1)

 Mixed (PaCO2>40 mmHg and HCO3-<24 40 (51.9) 9 (36) 31 (59.6)

mmol.l-1)

Lactic acidosis, n (%) < 0.001**

 No 11 (14.3) 8 (32) 3 (5.8)

 A-Type 57 (74) 10 (40) 47 (90.4)

 B-Type 9 (11.7) 7 (28) 2 (3.8)

Blood arterial lactate (mmol.l-1) 15.8 (9.9-20.5) 12.7 (3.4-18.9) 17 (11.4-22.3) 0.08

PaO2 (mm Hg) 143 (101-210) 134 (117-191) 145 (96-235) 0.93
-1
Base excess (mmol.l ) -23 (-26, -19) -24 (-26, -20) -23 (-26, -19) 0.84

14
 ACCEPTED MANUSCRIPT

SaO2 (%) 96 (89.8-98.3) 96 (91-98) 96 (84-99) 0.62

HCO3- (mEq.l-1) 8.6 (6.1, 13.4) 7.3 (5.9-12) 9.2 (6.4-13.5) 0.69
-1
Anionic gap (mEq.l ) 29 (23-37.3) 30.3 (20-40.6) 27.9 (23.4-33.4) 0.36

Serum creatinine (micromol.l-1) 178 (137-340) 347 (140-438) 166 (135-237) 0.01

Troponine Ic (ng.ml-1) 0.64 (0.22-2.22) 0.74 (0.24-2.98) 0.65 (0.17-1.6) 0.45

T
Prothrombin activity (%) 38 (17-56) 50 (28-65) 28.5 (16-46.3) 0.01

P
Na+ (mEq.l-1) 138 (132-145) 137 (132-142) 139 (134-145.5) 0.29

RI
+ -1
K (mEq.l ) 5.15 (4.1-5.9) 5.1 (4.6-6.1) 5.2 (4-5.8) 0.32

Cl- (mEql.l-1) 99 (93-106.3) 96 (92-102) 102 (93.5-108) 0.17

SC
AST (UI.l-1) 313 (75-1029) 119 (48-1410) 330 (109-858) 0.17

ALT (UI.l-1) 139 (34-381) 67 (30-597) 198 (49-360) 0.28

NU
LVEF (%) 47.5 (30-60) 45 (30-60) 50 (32.5-60) 0.7

Time delay for pH ≥7.3 (hours)¤ 17 (10-22) 17 (9.5-22) 19 (13.5-25) 0.42

Renal replacement therapy, n (%) 39 (50.6) 17 (68) 23 (44.2) 0.06

MA
Bicarbonate intra venous, n (%) 26 (33.8) 6 (24) 20 (38.5) 0.3

Extra corporeal membrane oxygenation, n (%) 4 (5.2) 2 (8) 2 (3.8) 0.59

Epinephrine and/or norepinephrine, n (%) 74 (96.1) 23 (92) 51 (98.1) 0.24

ED

Mechanical ventilation, n (%) 74 (96.1) 23 (92) 51 (98.1) 0.24

Length of ICU stay (days) 1.1 (0.4-6) 8 (4-15) 0.5 (0.2-1.1) < 0.001

*metabolic acidosis versus others (respiratory and mixed), **A-Type Lactic acidosis versus others, £data from 21
PT

patients, §data from 23 patients, ¤data from 32 patients (25 versus 7). Results are expressed as median (first and
CE

third quartile) and numbers (proportions).

ALT: alanine transaminase; AST: aspartate aminotransferase; COPD: chronic obstructive pulmonary disease;
AC

ICU: intensive care unit, LVEF: left ventricular ejection fraction, PaO2: partial pressure of arterial oxygen, SaO2:

arterial oxygen saturation, SAPS II: simplified acute physiology score II.

15
 ACCEPTED MANUSCRIPT

Table 2 Reason for ICU admission

Total Alive Dead P

(n=77) (n=25) (n=52)

Reason for ICU admission 0.002

T
 Septic shock, n (%) 15 (19.5) 5 (20) 10 (66)

P
 Cardiogenic shock, n (%) 11 (14.3) 3 (12) 8 (73)

RI
 Hypoxic cardiac arrest, n (%) 9 (11.7) 0 (0) 9 (100)

 Metformine associated lactic acidosis, n (%) 7 (9.1) 5 (20) 2 (29)

SC
 Hemorrhagic shock, n (%) 7 (9.1) 1(4) 6 (86)

 Hypercapnic respiratory failure, n (%) 7 (9.1) 4 (16) 3 (43)

NU
 Mesenteric infarction, n (%) 5 (6.5) 0 (0) 5 (100)

 Diabetic ketoacidosis, n (%) 2 (2.6) 2 (8) 0 (0)

 Thrombotic microangiopathy, n (%) 2 (2.6) 2 (8) 0 (0)

MA
 Necrotizing pancreatitis, n (%) 2 (2.6) 0 (0) 2 (100)

 Thiamine deficiency, n (%) 2 (2.6) 2 (8) 0 (0)

 Others, n (%) 8 (10.4) 1 (4) 7 (87)

ED
PT
CE
AC

16
 ACCEPTED MANUSCRIPT

Table 3 Risk factors independently associated with death in ICU, assessed in multi-variate analysis.

Variables Adjusted Odds Ratio (CI 95%) P

Serum creatinine 0.992 (0.986-0.998) 0.012

T
SAPS II 1.054 (1.014-1.095) 0.008

P
A-Type lactic acidosis 7.93 (1.78-35.35) 0.007

RI
CI: confidence intervals.

SC
We included, in the model, significantly different parameters (p <0.05) after bi-variate comparison: diabetes

NU
mellitus, prothrombine activity, SAPS II, A-Type lactic acidosis and serum creatinine. The “cardiac arrest before

admission” variable was not included due to multicollinearity. Prothrombine activity, and diabetes mellitus were
MA
excluded from the equation (P=0.33, and P=0.78, respectively). The Hosmer/Lemeshow test showed a good

model calibration (P=0.66). The Nagelkerke and Cox/Snell R squares were respectively 0.56 and 0.4.
ED
PT
CE
AC

17
 ACCEPTED MANUSCRIPT

TP
RI
SC
NU
MA
ED

Figure 1
PT
CE
AC

18
 ACCEPTED MANUSCRIPT

TP
RI
SC
NU
MA
ED

Figure 2
PT
CE
AC

19