Severe metabolic acidosis after out-of-hospital cardiac arrest: risk factors and association with outcome

Among successfully resuscitated out-of-hospital cardiac arrest (OHCA) patients admitted to the intensive care unit (ICU) after return of spontaneous circulation (ROSC), the occurrence of a “post-cardiac arrest syndrome” is the leading cause of early mortality. This syndrome is related to multi-organ ischemic reperfusion injury and may lead to multi-organ failure in the early hours after resuscitation [1].

Acidosis is frequently observed in this context [2] and is often multifactorial, resulting from major metabolic disturbances (increased levels of blood lactate, phosphate, unmeasured anions, acute kidney injury) as well as respiratory disturbances. Metabolic acidosis has well-known deleterious effects including decrease of cardiac contractility [3], arterial vasodilatation [4], impairment of the inflammatory and immune response [5] leading to an increased risk of multi-organ failure. Accordingly, a strong association between the depth of acidosis and poor outcome has been previously demonstrated in critically ill patients [6, 7]. Even though intuitive in OHCA patients, data on the prevalence and the severity of metabolic acidosis at ICU admission are scarce. Moreover, data on the prognostic impact of metabolic acidosis in these patients and on the proportion of patients with severe metabolic acidosis discharged alive from ICU are lacking.

In the present study, we aimed to describe the outcome of successfully resuscitated OHCA patients with metabolic acidosis and to identify factors associated with severe acidosis and survival to ICU discharge.

A total of 899 resuscitated OHCA patients were admitted to the study hospital during the study period. We excluded 27 patients without ROSC at ICU admission, one patient with DNR order and 2 patients with no available arterial blood gases at ICU admission. Among the 869 remaining patients with acid–base status available, 43 patients did not have metabolic acidosis, and we included 826 patients in the analysis.

Association between base deficit, ICU mortality and cause of death

Overall, 543 (65.7%) patients died in ICU. ICU mortality rate increased across base deficit quartiles (39.1, 59.2, 76.3 and 88.3%, p for trend < 0.001). ICU mortality was 42.2 versus 68.4% in patients without and with acidemia (p < 0.001). After adjustment for confounders, we observed an association between base deficit and mortality at ICU discharge (odds ratio 1.76 [1.07, 2.91], 3.82 [2.20, 6.65], 5.13 [2.67, 9.88] for 5–9, 9–13 and > 13 categories with base deficit < 5 as reference category, respectively) (Table 3). Similar results were obtained in patients without prehospital bicarbonate infusion and without either chronic respiratory disease or chronic kidney disease. We observed a strong interaction between base deficit and lactate level (p for interaction = 0.004).

Variable	Odds ratio	95% confidence interval	p value
Age (year)	1.03	1.02, 1.05	< 0.001
Male gender	1.39	0.87, 2.22	0.164
Public setting	0.63	0.41, 0.97	0.034
Witnessed CA	0.96	0.39, 2.35	0.931
Bystander CPR	0.76	0.50, 1.16	0.200
Initial VF/VT	0.43	0.26, 0.71	0.001
Collapse to ROSC, per min	1.05	1.03, 1.07	< 0.001
Cardiac cause-related CA	0.48	0.29, 0.80	0.005
Post-resus. shock	1.31	0.89, 1.96	0.181
Base deficit quartiles (mEq/l)
< 5	1.00	1.00, 1.00
5–9	1.76	1.07, 2.91	0.026
9–13	3.82	2.20, 6.65	< 0.001
> 13	5.13	2.67, 9.88	< 0.001

CA cardiac arrest; CPR cardiopulmonary resuscitation; ROSC restoration of spontaneous circulation

Brain damage was the leading cause of death (n = 367, 67.6%), while refractory shock was responsible for 173 (31.9%) deaths. Three patients died in ICU from septic shock without brain anoxic injury. Death related to refractory shock was more frequent in the subgroup of patients with the most severe metabolic acidosis (13.6, 22.1, 28.5 and 49.5% across base deficit quartiles). Overall, the proportion of CPC1 patients decreased across the base deficit quartile while it was similar among patients discharged alive (72.8, 67.1, 70.5 and 62.5%, p = 0.21) (Fig. 1). Among patients with the most severe acidosis, clinical characteristics except initial rhythm and in-hospital management were similar between patients who died and those who were discharged alive from ICU (Additional file 2: Table S1).

No patient survived with a base deficit higher than 25 mmol/l, a pH lower than 6.72, a blood bicarbonate level lower than 8.6 mmol/l or an admission arterial blood lactate level higher than 20 mmol/l (Additional file 3: Table S2).

In the present study, we found that metabolic acidosis was almost constantly observed at ICU admission in OHCA patients. Metabolic acidosis was related to the initial rhythm and the time interval from collapse to ROSC. Furthermore, we observed that metabolic acidosis was associated with ICU mortality in a severity-dependent manner and was also associated with refractory shock. However, we also found that about 10% of patients with a very severe metabolic acidosis (median base deficit > 13.2 mEq/l) survived to ICU discharge with good neurological recovery.

In the present study, metabolic acidosis was severe and strongly determined by the initial cardiac rhythm and the time interval from collapse to ROSC. While several studies have already reported the occurrence of acidosis after cardiac arrest, the present study reports findings from a large cohort of OHCA patients using a reproducible and reliable definition of metabolic acidosis [15–18]. To the best of our knowledge, the present study is the largest study focusing on the prevalence and the prognostic impact of metabolic acidosis in successfully resuscitated OHCA patients. Moreover, several studies reported metabolic acidosis using either pH and/or blood lactate level [19, 20]. This probably includes some patients with exclusive respiratory acidosis, especially in countries where there is no use of advanced airway in the field, and simultaneously misses some patients with compensated metabolic acidosis. This is the reason why we used the base deficit calculation without using pH. Furthermore, the association we observed between severity of metabolic acidosis and both resuscitation duration and initial rhythm may reflect the goodness-of-fit of such a measure with the degree and the duration of inadequate perfusion [21]. Thus, we showed that lactate level was an important contributor of metabolic acidosis but not sufficient to exhaustively describe it [22, 23]. In the present study, acidosis strongly correlated with both lactate level and phosphorus level, highlighting the fact that metabolic acidosis after OHCA is not only related to increase of lactate level but also to accumulation of unmeasured anions partially because of frequent acute kidney injury. This additional finding may explain the association we observed between severity of metabolic acidosis and ICU mortality as acute kidney injury has been demonstrated as an independent marker of ICU mortality after out-of-hospital cardiac arrest [2, 24].

Our findings support a strong association between severity of metabolic acidosis at ICU admission after OHCA and ICU mortality. Interestingly, we were able to provide cause of death and observed that ICU mortality in the most severe patients included in the present study was mostly related to refractory multi-organ failure. The association between metabolic acidosis and ICU mortality has been reported in different subgroups of critically ill patients: trauma [25], sepsis [26] as well as cardiac arrest [27–29]. Nolan et al. [27] reported a significant increase of 45% of in-hospital mortality per 0.1 point of pH decrease below 7.25. In the same manner, Chien et al. [29] reported an odds ratio of survival to hospital discharge of 10 for patients with a pH higher than 7.07 [95% confidence interval 2.1, 47.7]. Although pH is a simple measure to obtain at ICU admission, it does not distinguish metabolic and respiratory acidosis. Indeed, PCO₂ increase after ROSC may largely contribute to acidemia [30, 31]. Thus, we chose to consider base deficit, more likely to reflect metabolic acidosis and ischemia. We evidenced a severity-manner relationship between base deficit and survival to hospital discharge. Physicians may be aware that pH may add information to base deficit, as we observed that patients with acidemia had worse prognosis than those without. These findings are fully consistent with those previously published in some smaller cohorts [17, 28, 32, 33]. Besides the correlation between base deficit and duration of resuscitation efforts, metabolic acidosis may be associated with mortality by its hemodynamic effects, as depression of ventricular function, catecholamine release and reduction of ventricular responsiveness to catecholamines [4]. Interestingly, we did not evidence any relationship between metabolic acidosis and myocardial function, assessed by the left ventricle ejection function. This might suggest that in our cohort, metabolic acidosis may worsen vasoplegia in a more intensive manner and explain the increase of deaths related to refractory shock we observed across the base deficit quartiles.

We observed a survival rate of about 10% in the fourth quartile of base deficit, i.e., in patients with the most severe metabolic acidosis. In 1985, Weil et al. did not observe any survivor with a pH below 7.25 and an arterial lactate level higher than 7 mmol/l. However, this study was published 30 years ago and in-hospital management of successfully resuscitated OHCA patients has greatly improved since this publication [34]. In the study reported by Chien et al., the lowest pH associated with survival was 6.86 while Nolan et al. reported a survival rate of less than 1% below 6.70, about 3% between 6.70 and 6.80 and about 7% between 6.80 and 6.90 [27, 29]. Recently, Ilicki et al. [35] described 6 OHCA survivors with an initial pH lower than 6.90. However, among these cases, 3 had another underlying cause of acidemia (either ketoacidosis or metformin intoxication), and 2 had hemorrhagic shock. The comparison according to vital status at ICU discharge in this subgroup of patients did not evidence any in-hospital factors that could be associated with survival. We were particularly interested in the similar rate of renal replacement therapy in these patients regardless of the vital status at ICU discharge. Despite acidosis severity (almost three quarters of patients received renal replacement therapy within the first 24 h), several patients survived without aggressive treatment of metabolic acidosis and we did not find any additional factor that could differentiate these patients. Moreover, prehospital infusion of bicarbonate appeared to be strongly correlated with ICU mortality and seemed to be a surrogate marker of the severity assessed by the prehospital team.

We acknowledge several limitations in the present study. First, due to the retrospective design, we are not able to confirm the causality of the association we observed. Despite adjustment for prehospital and in-hospital confounders, we cannot be sure that metabolic acidosis might only be a mediator rather than an independent predictor of ICU mortality. Second, we specifically collected information about acidosis of OHCA patients included in analysis previously performed that focused on the prognostic impact of acute kidney injury in this setting [3]. Thus, we were not able to extend the study period beyond 2012. However, we do believe the results we observed would be similar in a more recent cohort. Third, we were not able to collect additional information on the pathogenesis of the metabolic acidosis. We did not collect chloremia, albumin or magnesium level, which could have allowed us to explore more deeply metabolic acidosis mechanisms. Fourth, we did not collect repeated measurements of base deficit, which prevented us to evaluate the prognostic impact of metabolic acidosis correction or acidosis duration. Last, we did not collect renal replacement therapy details, which could partly explain the differences between survivors and patients who died in the fourth quartile of base deficit.

Metabolic acidosis was observed in 95% of OHCA patients admitted to the ICU. Initial rhythm and time interval from collapse to ROSC were independently associated with base deficit. Base deficit was strongly associated with ICU mortality especially from refractory shock but seems to have no impact on neurological recovery in patients who were discharged alive. A substantial proportion of patients admitted with a very severe metabolic acidosis were discharged alive with a good neurological performance.