Dynamic changes in lactate and blood acid–base status during cardiopulmonary resuscitation and clinical outcomes in adult cardiac arrest patients: a prospective study Authors: Winchana Srivilaithon, Nutchapon Srikosai, Kumpol Kornthatchapong, Yaowapha Siripakarn Published: 2026 ABSTRACT Graphical abstract Introduction Cardiac arrest remains a major global public health challenge, with poor survival despite advances in resuscitation science.1 Although several prognostic models for return of spontaneous circulation (ROSC) have been proposed, most rely on subjective or pre-arrest variables rather than real-time physiological responses during cardio-pulmonary resuscitation (CPR).2 Current CPR guidelines recommend blood analysis during CPR as an adjunct to identify potentially reversible causes of cardiac arrest.3, 4 Among these parameters, lactate has attracted interest as a potential prognostic marker, but studies of initial lactate levels have shown inconsistent associations with ROSC.5, 6, 7, 8, 9, 10, 11, 12, 13, 14 During cardiac arrest, severe global ischemia leads to anaerobic metabolism, resulting in rising lactate levels and falling blood pH.11, 12, 13, 14, 15 Only two studies have evaluated dynamic lactate changes during CPR, with conflicting results.5, 6 Both were limited by small sample sizes and lacked a priori sample size estimation. A study from South Korea assessing dynamic arterial blood gas changes during CPR found that delta PaO2, but not lactate or pH, was associated with sustained ROSC.10 Hyperlactataemia and metabolic acidosis are associated with worse outcomes in critically ill patients and are often considered therapeutic targets.11, 16 During CPR, a reduction in lactate may reflect improved perfusion and ventilation, and lactate clearance may therefore be associated with better outcomes.17 However, lactate may also remain elevated or increase despite partial reperfusion because of ongoing buffering of excess hydrogen ions, while recovery of mitochondrial function may gradually improve acid–base status.15 In addition, impaired circulation during CPR may reduce hepatic lactate clearance, making the combined changes in lactate and blood pH potentially informative markers of resuscitation.18, 19 Given the conflicting evidence and uncertainty about the underlying physiology, we hypothesised that patients who achieved sustained ROSC would show more favourable metabolic changes during CPR, including greater lactate clearance and improvement in blood pH. Therefore, this study aimed to examine whether dynamic changes in lactate and blood acid–base status during CPR were associated with sustained ROSC, survival to hospital discharge, and favourable neurological outcomes. Methods Study setting and population This prospective cohort study was conducted in the emergency department (ED) of Thammasat University Hospital (TUH), a tertiary-care medical centre in Pathum Thani, Thailand. The ED manages approximately 50,000 patient visits annually and treats an estimated 120–150 cardiac arrest cases each year. Cardiac arrest patients are managed through an integrated emergency care system involving both the TUH emergency medical services (EMS) team and the ED resuscitation team, which are composed of the same group of emergency physicians working across both settings. The study included adult patients aged 18 years or older with cardiac arrest who were managed by the TUH EMS or ED resuscitation team and underwent point-of-care blood testing (POCT) on at least two occasions during resuscitation. The overall cohort comprised patients from four clinical settings: out-of-hospital cardiac arrest (OHCA) with resuscitation terminated at the scene, OHCA with CPR initiated at the scene and continued to the ED, OHCA presenting to the ED, and first cardiac arrest occurring in the ED. All patients were managed by emergency physicians and emergency medicine residents with 24-hour coverage, in accordance with internationally accepted standard treatment guidelines.3, 4 Patients were excluded if they had a do-not-resuscitate (DNR) order, showed obvious signs of death before CPR initiation, or if the patient or their legally authorised representative declined participation in the study. Research protocol This prospective cohort study was designed to evaluate whether dynamic changes in lactate levels and blood pH predict clinical outcomes in adult cardiac arrest patients managed by EMS or presenting to the ED. Venous blood samples were obtained through a peripheral venous cannula immediately after CPR initiation by a registered nurse or an emergency physician, with ultrasound assistance used for venous access when necessary. Samples were analysed using a POCT system (epoc® Blood Analysis System, Siemens Healthineers, Canada), which was available both in ambulances for prehospital EMS care and in the ED. The first sample was collected as soon as possible after recognition of cardiac arrest. POCT results were typically available within approximately 5 min, regardless of the resuscitation location. Resuscitation care and medication administration were provided by emergency physicians in accordance with standard treatment guidelines. Subsequent peripheral venous blood samples were collected at 10-min intervals according to a predefined protocol, or as soon as feasible. Repeated sampling continued until sustained ROSC was achieved, resuscitation was terminated, or the resuscitation team determined that further POCT sampling would provide no additional clinical benefit. In our ED, there was no protocol for central venous catheter insertion during ongoing CPR, and blood samples obtained from a central venous catheter or intraosseous cannulation were not used for POCT analysis. Serial blood sampling was part of the standard diagnostic evaluation for cardiac arrest patients in this setting and did not constitute an additional research-related procedure. This study involved data collection only and did not interfere with the resuscitation process or alter standard treatment protocols. Ethical considerations The study was approved by the Human Research Ethics Committee of the Faculty of Medicine, Thammasat University (MTU-EC-EM-0-191/67). Because cardiac arrest is a sudden, life-threatening emergency requiring immediate CPR, written informed consent could not be obtained before enrolment. A deferred consent process was therefore used. Written informed consent for continued data collection was obtained as soon as practicable from the patient or, when the patient lacked decision-making capacity, from the patient’s legally authorised representative. This manuscript was prepared in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. Data collection and outcomes Data were obtained from a prospectively maintained cardiac arrest registry routinely used at TUH. For all cardiac arrest cases, data collection began in the prehospital setting or in the ED and continued throughout resuscitation. Collected variables included demographic data, clinical characteristics of the arrest, resuscitation details, medications administered during CPR, POCT laboratory results (pH, PvCO2, HCO3, lactate, potassium, and glucose), and clinical outcomes. Each patient was included only once at the time of the first cardiac arrest event; patients with recurrent cardiac arrest after sustained ROSC were not re-enrolled or re-analysed. The primary outcome was sustained ROSC. Secondary outcomes were survival to hospital discharge or at least 30 days after cardiac arrest, and favourable neurological outcome, defined as a Cerebral Performance Category (CPC) score of 1 or 2 at discharge or 30 days.3, 20 Sample size calculation Data from a pilot study of cardiac arrest patients conducted in the ED of TUH between March and December 2022, in which lactate and blood pH were measured twice during CPR, were used to estimate the required sample size. The sample size was calculated using a two-sample comparison of means, with 90% power and a two-sided alpha of 0.05. Based on mean lactate levels of 11.5 ± 4.5 mmol/L at the initial measurement and 13.7 ± 3.9 mmol/L at the subsequent measurement, the estimated minimum sample size was 94 patients. Statistical analysis All data were recorded in a computerised database and analysed using STATA 17 (StataCorp LLC, College Station, TX, USA). Categorical variables are presented as frequencies and percentages, whereas continuous variables are presented as mean with standard deviation (SD) or median with interquartile range (IQR), as appropriate based on data distribution. Comparisons between patients with sustained ROSC and those without sustained ROSC were performed using the chi-square test or Fisher’s exact test for categorical variables, and the student’s t test or Wilcoxon rank-sum test for continuous variables, as appropriate. Changes in lactate levels and blood pH between the initial and subsequent POCT measurements were analysed using mixed-effects regression models for repeated measures, with results expressed as mean differences and 95% confidence intervals (CIs), including interaction terms. Associations between changes in lactate levels and blood pH and clinical outcomes were evaluated using multivariable logistic regression analyses. Covariates included in the adjusted models were selected based on clinical relevance and variables with a p-value < 0.05 identified from the univariable analysis. Results are reported as odds ratios (ORs) with 95% CIs. A two-sided p-value of <0.05 was considered statistically significant. Results During the 22-month study period, 302 patients with cardiac arrest were identified. Of these, 67 were excluded because they were younger than 18 years, had a do-not-resuscitate order, or did not undergo POCT measurement. An additional 72 patients were excluded because only a single POCT measurement was available, with no subsequent measurement obtained according to the study protocol. The final analysis therefore included 160 patients (Fig. 1). Sustained ROSC was achieved in 65 patients (40.6%). The median number of POCT measurements per patient was 2 (IQR 2–3). Overall, 99 patients (61.9%) underwent two POCT measurements, while the remaining 61 patients (38.1%) underwent at least three measurements.Fig. 1Flow of the study.CPR: Cardiopulmonary resuscitation; ROSC: Return of spontaneous circulation; POCT: point-of-care testing. Flow of the study. CPR: Cardiopulmonary resuscitation; ROSC: Return of spontaneous circulation; POCT: point-of-care testing. Baseline characteristics are presented in Table 1. Patients were predominantly male, with a mean age of approximately 60 years, and age and sex were similar between the sustained ROSC and non-sustained ROSC groups. Most cardiac arrests occurred out of hospital, particularly in the non-sustained ROSC group. The aetiology differed significantly between groups, with hypoxia more common in the sustained ROSC group and traumatic causes and hyperkalaemia more frequent in the non-sustained ROSC group. Most arrests were witnessed, whereas lay rescuer CPR and defibrillation were uncommon and similar between groups. Initial rhythms were predominantly non-shockable. Patients with sustained ROSC had a significantly shorter time from collapse to first chest compression. There were no significant differences between groups in the time from collapse to the first POCT measurement or in the number of POCT measurements obtained. Subgroup analyses presenting the baseline characteristics of patients with OHCA and EDCA are provided in Supplementary Tables S1 and S2, respectively. In addition, the baseline characteristics of patients excluded because only a single POCT measurement was obtained are presented in Supplementary Table S3.Table 1Baseline characteristics according to sustained ROSC in the overall cohort.CharacteristicTotal(N = 160)n (%)Sustained ROSC(N = 65)n (%)No Sustained ROSC(N = 95)n (%)P valueMale sex105 (65.6)46 (70.8)59 (62.1)0.25Age, mean (SD) (years)59.6 (19.5)57.6 (20.1)60.8 (19.2)0.28Initial location of cardiac arrest0.05 Out-of-hospital132 (82.5)49 (75.4)83 (87.4) EDCA28 (17.5)16 (24.6)12 (12.6)Location of resuscitation*0.005 OHCA; prehospital only11 (6.9)1 (1.5)10 (10.5) OHCA; prehospital then ED60 (37.5)30 (46.2)30 (31.6) OHCA; ED only61 (38.1)18 (27.7)43 (45.3) EDCA28 (17.5)16 (24.6)12 (12.6)Aetiology of cardiac arrest<0.001 Presume cardiac cause58 (36.3)20 (30.8)38 (40.0) Traumatic-related10 (6.3)010 (10.4) Hypoxia21 (13.1)16 (24.6)5 (5.3) Non-traumatic massive bleeding12 (7.5)7 (10.8)5 (5.3) Intoxication1 (0.6)1 (1.5)0 Sepsis related17 (10.6)10 (15.4)7 (7.4) Metabolic acidosis23 (14.4)8 (12.3)15 (15.8) Hyperkalaemia18 (11.2)3 (4.6)15 (15.8)Transfer methods0.009 Emergency medical service71 (44.4)34 (52.3)37 (38.9) First responder13 (8.1)1 (1.5)12 (12.6) Self-transport70 (43.8)30 (46.2)40 (42.2) Other method6 (3.7)06 (6.3)Witness cardiac arrest116 (72.5)51 (78.5)65 (68.4)0.16Lay rescuer CPR (n = 132)35 (26.5)16 (32.7)19 (22.9)0.22Lay rescuer defibrillation (n = 132)3 (2.3)2 (4.1)1 (1.2)0.28Initial cardiac rhythm0.64 Ventricular fibrillation21 (13.1)10 (15.4)11 (11.6) Pulseless ventricular tachycardia5 (3.1)2 (3.1)3 (3.2) Asystole66 (41.3)23 (35.4)43 (45.2) Pulseless electrical activity68 (42.5)30 (46.1)30 (40.0)Hypoglycemia at presentation15 (9.5)5 (7.7)10 (10.8)0.52EMS response time, mean (SD) (minute) (n = 71)13.8 (4.3)13.3 (3.6)14.2 (4.7)0.35Scene time, mean (SD) (minute) (n = 71)20.2 (9.9)16.8 (6.8)22.9 (11.1)0.009Time from collapse to first chest compression, median (IQR) (minute)11.5 (0, 30)5 (0, 25)15 (1, 34)0.02Time from collapse to first POCT measurement, median (IQR) (minute)23.5 (14, 59)22 (9, 54)25 (16, 60)0.11Time from advance life support initiation to first POCT measurement, median (IQR) (minute)5 (3, 7)5 (3, 7)5 (3, 8)0.10Number of POCT tests, median (IQR) (time)2 (2, 3)2 (2, 3)2 (2, 3)0.10Interval between first and second POCT, median (IQR) (minute)13 (10, 18.5)12 (10, 20)13 (10, 17)0.69 Initial POCT findings pH, mean (SD)6.95 (0.21)6.94 (0.19)6.96 (0.22)0.56 Lactate, mean (SD) (mmol/L)13.2 (4.8)13.0 (4.7)13.3 (4.9)0.71 PvCO2, mean (SD) (mmHg)67.7 (35.5)65.9 (33.2)68.9 (37.1)0.59 Potassium, mean (SD) (mmol/L)6.1 (2.1)5.7 (2.1)6.3 (2.1)0.07 HCO3, mean (SD) (mmol/L)15.5 (8.1)15.1 (10.3)15.7 (6.1)0.61ROSC: return of spontaneous circulation; EDCA: emergency department cardiac arrest; OHCA: out-of-hospital cardiac arrest; CPR: cardio-pulmonary resuscitation; EMS: emergency medical service; POCT: point-of-care testing; PvCO2: venous partial pressure of carbon dioxide; HCO3: bicarbonate.*OHCA; prehospital only = resuscitation was initiated and terminated in the prehospital setting. OHCA; prehospital then ED = resuscitation was initiated in the prehospital setting and continued after arrival in the ED. OHCA; ED only = OHCA in which resuscitation was first initiated after arrival in the ED. EDCA = cardiac arrest occurring in the ED. Patients with recurrent cardiac arrest after sustained ROSC were not re-enrolled as separate participants. Table 1Baseline characteristics according to sustained ROSC in the overall cohort.CharacteristicTotal(N = 160)n (%)Sustained ROSC(N = 65)n (%)No Sustained ROSC(N = 95)n (%)P valueMale sex105 (65.6)46 (70.8)59 (62.1)0.25Age, mean (SD) (years)59.6 (19.5)57.6 (20.1)60.8 (19.2)0.28Initial location of cardiac arrest0.05 Out-of-hospital132 (82.5)49 (75.4)83 (87.4) EDCA28 (17.5)16 (24.6)12 (12.6)Location of resuscitation*0.005 OHCA; prehospital only11 (6.9)1 (1.5)10 (10.5) OHCA; prehospital then ED60 (37.5)30 (46.2)30 (31.6) OHCA; ED only61 (38.1)18 (27.7)43 (45.3) EDCA28 (17.5)16 (24.6)12 (12.6)Aetiology of cardiac arrest<0.001 Presume cardiac cause58 (36.3)20 (30.8)38 (40.0) Traumatic-related10 (6.3)010 (10.4) Hypoxia21 (13.1)16 (24.6)5 (5.3) Non-traumatic massive bleeding12 (7.5)7 (10.8)5 (5.3) Intoxication1 (0.6)1 (1.5)0 Sepsis related17 (10.6)10 (15.4)7 (7.4) Metabolic acidosis23 (14.4)8 (12.3)15 (15.8) Hyperkalaemia18 (11.2)3 (4.6)15 (15.8)Transfer methods0.009 Emergency medical service71 (44.4)34 (52.3)37 (38.9) First responder13 (8.1)1 (1.5)12 (12.6) Self-transport70 (43.8)30 (46.2)40 (42.2) Other method6 (3.7)06 (6.3)Witness cardiac arrest116 (72.5)51 (78.5)65 (68.4)0.16Lay rescuer CPR (n = 132)35 (26.5)16 (32.7)19 (22.9)0.22Lay rescuer defibrillation (n = 132)3 (2.3)2 (4.1)1 (1.2)0.28Initial cardiac rhythm0.64 Ventricular fibrillation21 (13.1)10 (15.4)11 (11.6) Pulseless ventricular tachycardia5 (3.1)2 (3.1)3 (3.2) Asystole66 (41.3)23 (35.4)43 (45.2) Pulseless electrical activity68 (42.5)30 (46.1)30 (40.0)Hypoglycemia at presentation15 (9.5)5 (7.7)10 (10.8)0.52EMS response time, mean (SD) (minute) (n = 71)13.8 (4.3)13.3 (3.6)14.2 (4.7)0.35Scene time, mean (SD) (minute) (n = 71)20.2 (9.9)16.8 (6.8)22.9 (11.1)0.009Time from collapse to first chest compression, median (IQR) (minute)11.5 (0, 30)5 (0, 25)15 (1, 34)0.02Time from collapse to first POCT measurement, median (IQR) (minute)23.5 (14, 59)22 (9, 54)25 (16, 60)0.11Time from advance life support initiation to first POCT measurement, median (IQR) (minute)5 (3, 7)5 (3, 7)5 (3, 8)0.10Number of POCT tests, median (IQR) (time)2 (2, 3)2 (2, 3)2 (2, 3)0.10Interval between first and second POCT, median (IQR) (minute)13 (10, 18.5)12 (10, 20)13 (10, 17)0.69 Initial POCT findings pH, mean (SD)6.95 (0.21)6.94 (0.19)6.96 (0.22)0.56 Lactate, mean (SD) (mmol/L)13.2 (4.8)13.0 (4.7)13.3 (4.9)0.71 PvCO2, mean (SD) (mmHg)67.7 (35.5)65.9 (33.2)68.9 (37.1)0.59 Potassium, mean (SD) (mmol/L)6.1 (2.1)5.7 (2.1)6.3 (2.1)0.07 HCO3, mean (SD) (mmol/L)15.5 (8.1)15.1 (10.3)15.7 (6.1)0.61ROSC: return of spontaneous circulation; EDCA: emergency department cardiac arrest; OHCA: out-of-hospital cardiac arrest; CPR: cardio-pulmonary resuscitation; EMS: emergency medical service; POCT: point-of-care testing; PvCO2: venous partial pressure of carbon dioxide; HCO3: bicarbonate.*OHCA; prehospital only = resuscitation was initiated and terminated in the prehospital setting. OHCA; prehospital then ED = resuscitation was initiated in the prehospital setting and continued after arrival in the ED. OHCA; ED only = OHCA in which resuscitation was first initiated after arrival in the ED. EDCA = cardiac arrest occurring in the ED. Patients with recurrent cardiac arrest after sustained ROSC were not re-enrolled as separate participants. Baseline characteristics according to sustained ROSC in the overall cohort. ROSC: return of spontaneous circulation; EDCA: emergency department cardiac arrest; OHCA: out-of-hospital cardiac arrest; CPR: cardio-pulmonary resuscitation; EMS: emergency medical service; POCT: point-of-care testing; PvCO2: venous partial pressure of carbon dioxide; HCO3: bicarbonate. OHCA; prehospital only = resuscitation was initiated and terminated in the prehospital setting. OHCA; prehospital then ED = resuscitation was initiated in the prehospital setting and continued after arrival in the ED. OHCA; ED only = OHCA in which resuscitation was first initiated after arrival in the ED. EDCA = cardiac arrest occurring in the ED. Patients with recurrent cardiac arrest after sustained ROSC were not re-enrolled as separate participants. Management characteristics and outcomes are summarised in Table 2. Patients with sustained ROSC had a shorter time to first epinephrine administration. End-tidal carbon dioxide levels during CPR were significantly higher in the sustained ROSC group, whereas total epinephrine dose and CPR duration were significantly lower. Other resuscitation characteristics, including defibrillation strategies, airway management, and use of most medications, were similar between groups. Among patients who achieved sustained ROSC, 30.8% survived to hospital discharge and 12.3% had a favourable neurological outcome.Table 2Resuscitation management during CPR and clinical outcomes of the overall cardiac arrest cohort.CharacteristicTotal(N = 160)n (%)Sustained ROSC(N = 65)n (%)No Sustained ROSC(N = 95)n (%)P valueDefibrillation*Defibrillation during CPR71 (44.4)26 (40.0)45 (47.4)0.36Time to first defibrillation, median (IQR) (minute)4 (0, 11)3 (0, 7)5 (1, 13)0.13Total number of defibrillations, median (IQR) (time)2 (1, 5)1 (0, 4.5)3 (1, 6)0.08Defibrillation method0.93 Antero-lateral only50 (70.4)19 (73.1)31 (68.9) Double sequential external defibrillation18 (25.4)6 (23.1)12 (26.7) Vector change3 (4.2)1 (3.8)2 (4.4)Refractory shockable cardiac rhythm23 (14.4)8 (12.3)15 (15.8)0.54 AirwayAdvanced airway management0.39 Endotracheal intubation during CPR142 (88.7)55 (84.6)87 (91.6) Preexisting tracheostomy15 (9.4)9 (13.9)6 (6.3) Supraglottic airway during CPR3 (1.9)1 (1.5)2 (2.1)Time to advanced airway, median (IQR) (minute)2 (1, 3)2 (1, 3)2 (1, 3)0.97End-tidal carbon dioxide during CPR, median (IQR) (mmHg)25 (16, 33)30.5 (22, 40)20 (14, 30)<0.001 MedicineTime to first dose of epinephrine, median (IQR) (minute)2 (1, 3)1 (0, 2.5)2 (1, 3)0.03Total epinephrine dose, median (IQR) (milligramme)9 (6, 12)5 (3, 9)10 (9, 12)<0.001Epinephrine dosing period, median (IQR) (minute)3.3 (2.9, 4.7)3.4 (2.8, 5.5)3.2 (2.9, 3.9)0.38Intravenous amiodarone used during CPR40 (25.0)15 (23.1)25 (26.3)0.64Intravenous calcium used during CPR82 (51.6)27 (42.2)55 (57.9)0.05Intravenous sodium bicarbonate used during CPR78 (48.8)32 (49.2)46 (48.4)0.92Intravenous lidocaine used during CPR2 (1.3)1 (1.5)1 (1.1)0.79Intravenous insulin used during CPR47 (29.4)13 (20.0)34 (35.8)0.03Intravenous glucose used during CPR39 (24.4)14 (21.5)25 (26.3)0.49 OutcomeTotal CPR time, median (IQR) (minute)33 (24, 50)26 (10, 41)36 (30, 56)<0.001Coronary intervention7 (4.4)7 (10.8)NATemperature control with hypothermia9 (5.6)9 (13.9)NASurvival to hospital discharge20 (12.5)20 (30.8)NAFavourable neurological outcome8 (5.2)8 (12.3)NAROSC: return of spontaneous circulation; CPR: cardio-pulmonary resuscitation.*Defibrillation-related variables were calculated from the entire cohort and included both patients with an initial shockable rhythm and those who developed a shockable rhythm during CPR. Refractory shockable rhythm was defined as an initial shockable rhythm that persisted despite at least three defibrillation attempts. Table 2Resuscitation management during CPR and clinical outcomes of the overall cardiac arrest cohort.CharacteristicTotal(N = 160)n (%)Sustained ROSC(N = 65)n (%)No Sustained ROSC(N = 95)n (%)P valueDefibrillation*Defibrillation during CPR71 (44.4)26 (40.0)45 (47.4)0.36Time to first defibrillation, median (IQR) (minute)4 (0, 11)3 (0, 7)5 (1, 13)0.13Total number of defibrillations, median (IQR) (time)2 (1, 5)1 (0, 4.5)3 (1, 6)0.08Defibrillation method0.93 Antero-lateral only50 (70.4)19 (73.1)31 (68.9) Double sequential external defibrillation18 (25.4)6 (23.1)12 (26.7) Vector change3 (4.2)1 (3.8)2 (4.4)Refractory shockable cardiac rhythm23 (14.4)8 (12.3)15 (15.8)0.54 AirwayAdvanced airway management0.39 Endotracheal intubation during CPR142 (88.7)55 (84.6)87 (91.6) Preexisting tracheostomy15 (9.4)9 (13.9)6 (6.3) Supraglottic airway during CPR3 (1.9)1 (1.5)2 (2.1)Time to advanced airway, median (IQR) (minute)2 (1, 3)2 (1, 3)2 (1, 3)0.97End-tidal carbon dioxide during CPR, median (IQR) (mmHg)25 (16, 33)30.5 (22, 40)20 (14, 30)<0.001 MedicineTime to first dose of epinephrine, median (IQR) (minute)2 (1, 3)1 (0, 2.5)2 (1, 3)0.03Total epinephrine dose, median (IQR) (milligramme)9 (6, 12)5 (3, 9)10 (9, 12)<0.001Epinephrine dosing period, median (IQR) (minute)3.3 (2.9, 4.7)3.4 (2.8, 5.5)3.2 (2.9, 3.9)0.38Intravenous amiodarone used during CPR40 (25.0)15 (23.1)25 (26.3)0.64Intravenous calcium used during CPR82 (51.6)27 (42.2)55 (57.9)0.05Intravenous sodium bicarbonate used during CPR78 (48.8)32 (49.2)46 (48.4)0.92Intravenous lidocaine used during CPR2 (1.3)1 (1.5)1 (1.1)0.79Intravenous insulin used during CPR47 (29.4)13 (20.0)34 (35.8)0.03Intravenous glucose used during CPR39 (24.4)14 (21.5)25 (26.3)0.49 OutcomeTotal CPR time, median (IQR) (minute)33 (24, 50)26 (10, 41)36 (30, 56)<0.001Coronary intervention7 (4.4)7 (10.8)NATemperature control with hypothermia9 (5.6)9 (13.9)NASurvival to hospital discharge20 (12.5)20 (30.8)NAFavourable neurological outcome8 (5.2)8 (12.3)NAROSC: return of spontaneous circulation; CPR: cardio-pulmonary resuscitation.*Defibrillation-related variables were calculated from the entire cohort and included both patients with an initial shockable rhythm and those who developed a shockable rhythm during CPR. Refractory shockable rhythm was defined as an initial shockable rhythm that persisted despite at least three defibrillation attempts. Resuscitation management during CPR and clinical outcomes of the overall cardiac arrest cohort. ROSC: return of spontaneous circulation; CPR: cardio-pulmonary resuscitation. Defibrillation-related variables were calculated from the entire cohort and included both patients with an initial shockable rhythm and those who developed a shockable rhythm during CPR. Refractory shockable rhythm was defined as an initial shockable rhythm that persisted despite at least three defibrillation attempts. Results of the multilevel repeated-measures analysis comparing initial and follow-up POCT values between patients with and without sustained ROSC are presented in Table 3. In patients who achieved sustained ROSC, blood pH showed a small, non-significant increase over time, whereas pH significantly decreased in the non-sustained ROSC group. The interaction between time and ROSC status was significant (interaction P = 0.01), indicating divergent pH trajectories between groups. Lactate levels remained stable in the sustained ROSC group but increased significantly in the non-sustained ROSC group, with a significant interaction effect (interaction P = 0.001). PvCO2 increased significantly over time in the non-sustained ROSC group but not in the sustained ROSC group, although the interaction term was not statistically significant. Potassium levels decreased significantly over time in the sustained ROSC group but showed no significant change in the non-sustained ROSC group, with a significant interaction effect (interaction P = 0.03). No significant within-group changes or interaction effects were observed for bicarbonate (HCO3).Table 3Multilevel repeated-measures analysis of initial and follow-up point-of-care blood testing values, mean differences, and interaction effects by sustained ROSC status. ROSC: return of spontaneous circulation.ParameterSustained ROSC(N = 65)No Sustained ROSC(N = 95)Interaction*Initialmean (SD)Followingmean (SD)Initialmean (SD)Followingmean (SD)pH6.94 (0.19)6.98 (0.19)6.97 (0.22)6.91 (0.24) Mean difference0.04−0.060.09 95% confidence interval−0.01, 0.09−0.11, −0.010.02, 0.17 P value0.130.020.01 Lactate (mmol/L)13.0 (4.7)12.7 (4.7)13.3 (4.9)15.1 (4.4) Mean difference−0.271.80−2.07 95% confidence interval−1.22, 0.680.98, 2.62−3.31, −0.84 P value0.57<0.0010.001 PvCO2 (mmHg)65.9 (33.2)68.3 (31.8)68.9 (37.1)80.7 (40.5) Mean difference2.3911.80−9.41 95% confidence interval−7.21, 11.992.94, 20.66−22.46, 3.64 P value0.620.0090.16 Potassium (mmol/L)5.7 (2.1)4.9 (1.8)6.3 (2.1)6.3 (2.4) Mean difference−0.75−0.07−0.68 95% confidence interval−1.24, −0.27−0.47, 0.34−1.29, −0.06 P value0.0030.750.03 HCO3 (mmol/L)15.1 (10.4)15.7 (5.9)15.8 (6.3)16.8 (13.2) Mean difference0.620.99−0.41 95% confidence interval−2.25, 3.49−2.6, 4.6−4.69, 3.87 P value0.670.590.85*Interaction term derived from mixed-effects regression modelling of repeated measurements (initial vs. follow-up). Table 3Multilevel repeated-measures analysis of initial and follow-up point-of-care blood testing values, mean differences, and interaction effects by sustained ROSC status. ROSC: return of spontaneous circulation.ParameterSustained ROSC(N = 65)No Sustained ROSC(N = 95)Interaction*Initialmean (SD)Followingmean (SD)Initialmean (SD)Followingmean (SD)pH6.94 (0.19)6.98 (0.19)6.97 (0.22)6.91 (0.24) Mean difference0.04−0.060.09 95% confidence interval−0.01, 0.09−0.11, −0.010.02, 0.17 P value0.130.020.01 Lactate (mmol/L)13.0 (4.7)12.7 (4.7)13.3 (4.9)15.1 (4.4) Mean difference−0.271.80−2.07 95% confidence interval−1.22, 0.680.98, 2.62−3.31, −0.84 P value0.57<0.0010.001 PvCO2 (mmHg)65.9 (33.2)68.3 (31.8)68.9 (37.1)80.7 (40.5) Mean difference2.3911.80−9.41 95% confidence interval−7.21, 11.992.94, 20.66−22.46, 3.64 P value0.620.0090.16 Potassium (mmol/L)5.7 (2.1)4.9 (1.8)6.3 (2.1)6.3 (2.4) Mean difference−0.75−0.07−0.68 95% confidence interval−1.24, −0.27−0.47, 0.34−1.29, −0.06 P value0.0030.750.03 HCO3 (mmol/L)15.1 (10.4)15.7 (5.9)15.8 (6.3)16.8 (13.2) Mean difference0.620.99−0.41 95% confidence interval−2.25, 3.49−2.6, 4.6−4.69, 3.87 P value0.670.590.85*Interaction term derived from mixed-effects regression modelling of repeated measurements (initial vs. follow-up). Multilevel repeated-measures analysis of initial and follow-up point-of-care blood testing values, mean differences, and interaction effects by sustained ROSC status. ROSC: return of spontaneous circulation. Interaction term derived from mixed-effects regression modelling of repeated measurements (initial vs. follow-up). Table 4 presents the results of multivariable logistic regression analyses examining associations between changes in lactate, changes in pH, and clinical outcomes. After adjustment for relevant demographic, prehospital, and resuscitation variables, a greater decrease in lactate levels was independently associated with higher odds of sustained ROSC (adjusted OR 1.16 per 1 mmol/L decrease; 95% CI 1.05–1.28; P = 0.003) and survival to hospital discharge (adjusted OR 1.16; 95% CI 1.01–1.32; P = 0.03), but not with favourable neurological outcome. An increase in pH was also independently associated with sustained ROSC (adjusted OR 1.23 per 0.1-unit increase; 95% CI 1.03–1.48; P = 0.02), whereas associations with survival to hospital discharge and favourable neurological outcome did not reach statistical significance. Sensitivity analysis in patients with OHCA is presented in Supplementary Table S4. A greater decrease in lactate levels and an increase in pH were independently associated with higher odds of sustained ROSC. Positive trends were also observed for survival to hospital discharge and favourable neurological outcome; however, these associations were not statistically significant.Table 4Multivariable logistic regression analysis of delta lactate and delta pH for sustained ROSC, survival to hospital discharge, and favourable neurological outcome.Outcome*(N = 160)UnadjustedAdjusted**Odds ratio95% confidence intervalP valueOdds ratio95% confidence intervalP valueDelta lactate (per 1 mmol/L decrease in lactate levels) Sustained ROSC1.141.05, 1.250.0021.161.05, 1.280.003 Survival to discharge1.131.01, 1.270.021.161.01, 1.320.03 Favourable neurological outcome1.090.92, 1.290.301.320.98, 1.770.07 Delta pH (per 0.1-unit increase in pH) Sustained ROSC1.221.05, 1.430.011.231.03, 1.480.02 Survival to discharge1.130.94, 1.370.191.290.98, 1.700.07 Favourable neurological outcome1.150.89, 1.490.291.970.99, 3.890.05*For delta lactate, the results are presented per 1 mmol/L decrease in lactate levels. For delta pH, the results are presented per 0.1-unit increase in pH.**Adjusted with sex, age, initial location of cardiac arrest, location of resuscitation, aetiology of cardiac arrest, transfer method, time from collapse to first chest compression, initial cardiac rhythm, time to first dose of epinephrine, use of amiodarone, calcium, sodium bicarbonate, and insulin. Table 4Multivariable logistic regression analysis of delta lactate and delta pH for sustained ROSC, survival to hospital discharge, and favourable neurological outcome.Outcome*(N = 160)UnadjustedAdjusted**Odds ratio95% confidence intervalP valueOdds ratio95% confidence intervalP valueDelta lactate (per 1 mmol/L decrease in lactate levels) Sustained ROSC1.141.05, 1.250.0021.161.05, 1.280.003 Survival to discharge1.131.01, 1.270.021.161.01, 1.320.03 Favourable neurological outcome1.090.92, 1.290.301.320.98, 1.770.07 Delta pH (per 0.1-unit increase in pH) Sustained ROSC1.221.05, 1.430.011.231.03, 1.480.02 Survival to discharge1.130.94, 1.370.191.290.98, 1.700.07 Favourable neurological outcome1.150.89, 1.490.291.970.99, 3.890.05*For delta lactate, the results are presented per 1 mmol/L decrease in lactate levels. For delta pH, the results are presented per 0.1-unit increase in pH.**Adjusted with sex, age, initial location of cardiac arrest, location of resuscitation, aetiology of cardiac arrest, transfer method, time from collapse to first chest compression, initial cardiac rhythm, time to first dose of epinephrine, use of amiodarone, calcium, sodium bicarbonate, and insulin. Multivariable logistic regression analysis of delta lactate and delta pH for sustained ROSC, survival to hospital discharge, and favourable neurological outcome. For delta lactate, the results are presented per 1 mmol/L decrease in lactate levels. For delta pH, the results are presented per 0.1-unit increase in pH. Adjusted with sex, age, initial location of cardiac arrest, location of resuscitation, aetiology of cardiac arrest, transfer method, time from collapse to first chest compression, initial cardiac rhythm, time to first dose of epinephrine, use of amiodarone, calcium, sodium bicarbonate, and insulin. Fig. 2 illustrates serial changes in lactate and pH levels at each POCT measurement time point during CPR, stratified by sustained ROSC, survival to hospital discharge, and favourable neurological outcome. Overall, patients with favourable outcomes tended to show lower or decreasing lactate levels and gradual increases in pH over repeated measurements, whereas those without favourable outcomes generally showed persistently higher or increasing lactate levels and lower pH values over time.Fig. 2Serial changes in lactate and pH during cardiopulmonary resuscitation according to clinical outcomes.Boxplots showing (1) lactate level and (2) pH level at each measurement time point during CPR, stratified by sustained ROSC, survival to hospital discharge, and favourable neurological outcome. ROSC: Return of spontaneous circulation. Serial changes in lactate and pH during cardiopulmonary resuscitation according to clinical outcomes. Boxplots showing (1) lactate level and (2) pH level at each measurement time point during CPR, stratified by sustained ROSC, survival to hospital discharge, and favourable neurological outcome. ROSC: Return of spontaneous circulation. Discussion This prospective cohort study found that decreases in lactate levels and increases in pH during CPR, measured using serial peripheral venous POCT at approximately 10-min intervals, were associated with better resuscitation outcomes in adult patients with cardiac arrest, including sustained ROSC and survival to hospital discharge or 30 days. However, no significant association was observed with favourable neurological outcome, which may reflect limited statistical power. Previous studies by Contenti et al.5 and Vujanović et al.,6 which measured lactate levels during CPR, did not find a significant association between sustained ROSC and either the initial lactate level or changes in lactate over time. In contrast, our study found a significant association, which may be related to the larger sample size. Several other studies that measured lactate only once during CPR reported that lower initial lactate levels were associated with sustained ROSC.7, 8, 9, 10, 11, 12, 13, 14 However, in our cohort, initial lactate levels were not significantly different between patients with and without sustained ROSC. One possible explanation is that the resuscitation process itself may influence lactate levels. Effective CPR may improve perfusion and alter lactate dynamics during ongoing resuscitation, so serial changes in lactate may be more closely related to outcomes than the initial lactate value alone. Changes in lactate levels during CPR were significantly associated with sustained ROSC in our study. This finding may reflect more effective resuscitation, as high-quality CPR can improve blood flow and tissue perfusion, partially restore aerobic metabolism, reduce lactate production, and lessen metabolic acidosis in patients with cardiac arrest. Improved circulation during CPR may also enhance hepatic lactate clearance, contributing to a downward trend in lactate levels among patients who achieve sustained ROSC.5 Previous studies have examined whether different FiO2 levels during CPR affect lactate changes, but did not find a clear association between the level of FiO2 administered and lactate reduction.21, 22 In our study, all patients undergoing CPR received ventilation with 100% FiO2 in accordance with current international guidelines. Therefore, we believe that the observed changes in lactate metabolism and acid–base status were more likely to reflect the overall physiological response to resuscitation, including changes in tissue perfusion and the effectiveness of treatment directed at the underlying cause of cardiac arrest. Previous studies have consistently demonstrated strong associations between lactate levels and survival to hospital discharge, based on either initial lactate measurements obtained during CPR9, 14, 23 or lactate levels measured after ROSC.24, 25, 26, 27, 28, 29 However, studies evaluating dynamic lactate changes have largely focused on the post-ROSC period,24, 29, 30 whereas our study uniquely assessed dynamic lactate changes during ongoing CPR. Nevertheless, dynamic lactate changes at various time points have been shown to be associated with survival to hospital discharge.24, 29, 30 In contrast, changes in lactate levels during CPR were not associated with favourable neurological outcomes in our study. This finding differs from previous reports demonstrating significant associations when lactate was measured only at the initial time point9, 23, 24, 26, 31 or when dynamic lactate changes were assessed after ROSC.24, 29, 31, 32 Larger studies specifically evaluating lactate dynamics during CPR are required to further clarify their relationship with neurological outcomes. Regarding pH changes, several previous studies have demonstrated that a higher initial pH during CPR is associated with an increased likelihood of achieving sustained ROSC,7, 8, 11, 33 although some studies have reported no significant association.10, 12, 13 With respect to survival to hospital discharge or favourable neurological outcomes, higher initial pH values measured either during CPR23, 33, 34 or after ROSC25, 26 have shown strong associations. However, no prior studies have evaluated dynamic changes in pH during CPR in relation to these outcomes. To our knowledge, this study is the first to demonstrate an association between changes in pH during CPR and the likelihood of achieving sustained ROSC. The observed association between higher or increasing pH during CPR and sustained ROSC may reflect partial reversal of metabolic acidosis caused by anaerobic metabolism through more effective CPR, resulting in a trend toward alkalinization.18, 19 Although pH changes were associated with sustained ROSC in our study, larger sample sizes may be required to demonstrate significant associations with survival to hospital discharge or favourable neurological outcomes. In this study, dynamic changes in lactate levels and blood pH were associated with favourable clinical outcomes. We believe that the observed changes in lactate metabolism and acid–base status are more likely to reflect the overall physiological response to resuscitation, including changes in tissue perfusion and the effectiveness of treatment directed at the underlying cause of cardiac arrest. In this context, POCT may be valuable because it provides rapid bedside information on the patient’s metabolic status during resuscitation and may facilitate timely recognition of important physiological changes. We therefore interpret the delta changes in lactate and pH as potentially reflecting the patient’s response to ongoing resuscitative efforts and as supporting a more individualised treatment approach based on the evolving clinical condition of each patient. In addition, the findings of this study may provide a basis for further research on the use of these dynamic metabolic changes to predict patient outcomes during CPR. Future studies may focus on the development and validation of predictive models incorporating these dynamic metabolic parameters. Limitations This study has several limitations. First, although POCT sampling was planned at 10-min intervals during ongoing CPR, the actual interval was slightly longer, with a median of approximately 13 min. In addition, repeated measurements beyond the initial and second samples were not consistently available, as many patients achieved ROSC or had resuscitation terminated before further sampling could be performed. Therefore, the analysis of lactate and pH dynamics was based mainly on the most consistently available repeated measurements. Second, 72 patients were excluded from the main analysis because follow-up POCT measurements were unavailable. This may have introduced selection bias by excluding patients with either a very good or a very poor prognosis. Among these 72 patients, 47 (65.3%) achieved sustained ROSC before the next scheduled POCT measurement. Of the remaining 25 patients (34.7%) whose CPR duration was long enough that repeat sampling would have been expected, 14 (19.4% of the excluded group) had repeat samples sent but no result was available because of measurement error during analysis, while in 11 patients (15.3%) the treating physicians considered that further POCT sampling would provide no additional clinical benefit. Third, treatment during CPR was not standardised by the study protocol. All interventions were provided at the discretion of the treating physicians in accordance with current international standard guidelines. Consequently, POCT results may have influenced subsequent therapeutic decisions, which in turn could have affected later POCT measurements. The observed dynamic changes in lactate and pH may therefore reflect not only the patient’s physiological response during resuscitation but also the effects of ongoing treatment. In addition, we did not examine in detail which specific resuscitation interventions had the greatest impact on lactate and pH changes during CPR, limiting our ability to attribute these metabolic changes to individual components of resuscitation care. Finally, this was a single-centre study conducted at a tertiary-care medical centre in Thailand with advanced resuscitation capabilities and rapid access to POCT, which may limit the generalisability of the findings to other settings with fewer resources or less timely blood sample analysis. Finally, analyses of survival to hospital discharge and favourable neurological outcomes may have been underpowered because of the limited sample size, and further studies with larger cohorts are needed to confirm these findings. Conclusion Dynamic changes in metabolic parameters during CPR were associated with clinical outcomes in patients with cardiac arrest. Decreases in lactate levels and increases in pH measured by serial POCT during CPR were associated with higher rates of sustained ROSC and survival to hospital discharge. No significant associations were found with favourable neurological outcome. These findings suggest a potential association between serial metabolic changes during CPR and resuscitation outcomes, but further validation in larger, multicentre studies is required. CRediT authorship contribution statement Winchana Srivilaithon: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Nutchapon Srikosai: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Kumpol Kornthatchapong: Writing – review & editing, Supervision, Methodology, Conceptualization. Yaowapha Siripakarn: Writing – review & editing, Supervision, Methodology, Conceptualization. Funding This study was supported by the Research Fund of the Faculty of Medicine, 10.13039/501100005790Thammasat University (Fiscal Year 2025). Declaration of competing interest The authors have no conflicts of interest to disclose.