Kerala Surgical Journal

ORIGINAL ARTICLE
Year
: 2022  |  Volume : 28  |  Issue : 1  |  Page : 20--23

Arterial pH, bicarbonate levels and base deficit as predictors of morbidity and mortality in acute pancreatitis


Joicy Els Jojo, Satish G Prabhu, Vergis Paul, MG Jayan 
 Department of General Surgery, MOSC Medical College, Kolenchery, Kerala, India

Correspondence Address:
Dr. Joicy Els Jojo
Department of General Surgery, MOSC Medical College, Kolenchery - 682 311, Kerala
India

Abstract

Background: Although multiple criteria are available for prognosis and determination of severity of acute pancreatitis (AP), there is a need to identify better predictors of developing complications. Arterial pH is a component of the APACHE-II scoring system, commonly used in AP. Aim: The aim of this article is to study the diagnostic accuracy of arterial pH, bicarbonate levels and base deficit in predicting morbidity and mortality in AP. Materials and Methods: A cross-sectional study (agreement analysis) was performed on inpatients with AP in a tertiary hospital. One hundred and forty patients based on clinical, biochemical and radiological findings were included in this study. Patients with AP following trauma/surgical intervention, tumours of pancreas and patients with age ≤18 years were excluded all categorical variables were summarised using frequency and percentage. All quantitative variables were summarised using mean and standard deviation sensitivity, specificity, positive predictive value and negative predictive value were estimated. Statistical analysis was performed using EZ R software. Results: All participants who required ventilatory support, haemodialysis or blood product transfusion had a pH of <7.35. 12/13 (92.3%) of patients who were on haemodynamic supports had bicarbonate levels <22mmol/L. Those who required respiratory support had bicarbonate level <22 mmol/L and base deficit <–4 mmol/L. 11/13 (84.6%) who required haemodynamic supports had base deficit <–4 mmol/L. All 5 of the 140 (3.57%) participants who expired had an arterial pH <7.35, bicarbonate level <22 mmol/L and base deficit <–4 mmol/L. Conclusion: In patients with AP, low arterial pH, low bicarbonate levels and high base deficit at presentation predict an adverse outcome with a higher frequency of organ failure and mortality.



How to cite this article:
Jojo JE, Prabhu SG, Paul V, Jayan M G. Arterial pH, bicarbonate levels and base deficit as predictors of morbidity and mortality in acute pancreatitis.Kerala Surg J 2022;28:20-23


How to cite this URL:
Jojo JE, Prabhu SG, Paul V, Jayan M G. Arterial pH, bicarbonate levels and base deficit as predictors of morbidity and mortality in acute pancreatitis. Kerala Surg J [serial online] 2022 [cited 2022 Nov 30 ];28:20-23
Available from: http://www.keralasurgj.com/text.asp?2022/28/1/20/350887


Full Text



 Introduction



Acute pancreatitis (AP) is an acute inflammatory process of the pancreas. Although multiple criteria are available for prognosis and determination of severity of AP, there is a need to identify better predictors of developing complications. Predicting the severity and outcome of AP still represents a challenge for the physician. Although there are multiple scoring systems to predict a severe outcome in AP, there is no consensus regarding the use of one single marker/system or the other in clinical practice. This paper aims to study the diagnostic accuracy of arterial pH in predicting morbidity and mortality in AP and to study the diagnostic accuracy of arterial bicarbonate levels and arterial base deficit in predicting morbidity and mortality in AP.

 Materials and Methods



A cross-sectional study (agreement analysis) was performed on inpatients diagnosed with AP in a tertiary hospital. The sample size was 136.6, calculated based on sensitivity estimation. Inclusion criteria were all patients admitted with AP based on clinical, biochemical and radiological findings. Exclusion criteria included AP following trauma/surgical intervention, tumours of pancreas causing pancreatitis and patient with age ≤18 years. Data were collected using a structured pro forma after obtaining consent of all patients. All categorical variables were summarised using frequency and percentage. All quantitative variables were summarised using mean and standard deviation sensitivity, specificity, positive predictive value and negative predictive value were estimated. Statistical analysis was performed using EZ R software. Ethical approval was sought from institutional ethics committee.

 Results



The average age of the participants was 48.06 ± 18.04 years. The minimum and maximum age was 18 and 92 years, respectively [Figure 1].{Figure 1}

The majority of the study participants were male. 74.3% (104/140) study participants were males and 25.7% (36/140) females.

The most common clinical symptom at the time of presentation, 92.9% (130/140) presented with abdominal pain, and 29.2% (38/140) had the classical feature of pain radiating to the back [Table 1].{Table 1}

50.7% (71/140) patients did not have comorbidities. 49.3% (69/140) patients had comorbidities, with few suffering from multiple comorbidities [Table 2].{Table 2}

80% (112/140) of the participants had abdominal tenderness on clinical examination, whereas 20% (28/140) did not. The association between CT severity index and amylase at the time of admission was significant (P < 0.05). The amylase level at admission was significantly higher in severe cases (1798 ± 2366) and moderate cases (1503 ± 2026) compared to mild cases (521.0 ± 580.2).

There was a significant association between computed tomography severity and lipase level at admission (P < 0.05). Lipase level at admission was significantly higher in severe cases (12909 ± 12851) and moderate cases (11241 ± 11866) compared to mild cases (3524 ± 5138).

It was noted that none of the study participants required surgical intervention at any point of the duration in hospital.

Arterial pH was a significant predictor of mortality (P < 0.05). pH <7.35 was associated with higher rates of mortality. All 5/140 (3.57%) participants who expired had an arterial pH <7.35. The sensitivity of arterial pH as a predictor of mortality associated with AP was 100%. However, the specificity was only 40%. The positive predictive value was 6% and the negative predictive value was 100%. It was not a significant predictor of the need for haemodynamic supports. 11/13 (84.6%) participants who were on haemodynamic supports had an arterial pH of <7.35. The sensitivity of arterial pH as a predictor of haemodynamic supports was 85%, specificity 41%, positive predictive value 13% and negative predictive value 96%. It was a significant predictor of blood products transfusion (P < 0.05). All 19/140 (100%) participants required blood product transfusion with sensitivity of 100%, specificity of 45%, positive predictive value of 22% and negative predictive value of 100%. It was also a significant predictor of need for TPN (P < 0.05). All 19 participants who were given TPN during the course in hospital had an arterial pH <7.35. The sensitivity of arterial pH as a predictor of need for TPN is 100%, with 45% specificity, 22% positive predictive value and 100% negative predictive value. It was a significant predictor of need for ventilator (P < 0.05). All 10 participants who required respiratory support in the form of mechanical ventilation had arterial pH <7.35. The sensitivity of arterial pH as marker to predict the need for mechanical ventilation was 100% with specificity being 42%, positive predictive value of 12% and negative predictive value of 100%. It was a significant predictor of need for haemodialysis (P < 0.05). This indicates that it is a significant predictor of renal injury/failure. All 13 participants who required haemodialysis had arterial pH <7.35. The sensitivity of arterial pH as a predictor of need for haemodialysis was 100%, with a specificity of 43%, positive predictive value of 15% and negative predictive value of 100%. However, it was not a significant predictor of the duration of hospital stay. Arterial pH was not a significant predictor of the duration of intensive care unit (ICU) stay.

Arterial bicarbonate was a significant predictor of mortality (P < 0.005). Value <22 mmol/L was significantly associated with mortality. All 5 participants (3.57%) who expired had bicarbonate level <22 mmol/L. The sensitivity of arterial bicarbonate as a predictor of mortality associated with AP was 100%. The specificity of arterial bicarbonate as a predictor of mortality was only 49%. The positive predictive value was 7% and the negative predictive value was 100%. It was also a significant predictor of haemodynamic supports (P < 0.05). 12/13 (92.3%) participants who were on haemodynamic supports had arterial levels <22 mmol/L. The sensitivity of arterial bicarbonate as a predictor of need for haemodynamic supports was 92%, specificity 51%, positive predictive value 16% and negative predictive value 98%. It was a significant predictor of blood products transfusion (P < 0.05) with sensitivity of 89%, specificity of 53%, positive predictive value of 23% and negative predictive value of 97%. It was also a significant predictor of need for total parenteral nutrition (TPN) (P < 0.05). The sensitivity of the same being 89%, specificity being 53%, positive predictor value being 23% and negative predictive value being 97%. Arterial bicarbonate was a significant predictor of ventilator support (P < 0.05). All 10 participants who required respiratory support in the form of mechanical ventilation had arterial bicarbonate level <22 mmol/L. The sensitivity of arterial bicarbonate as a marker to predict the need for mechanical ventilation was 100% with specificity being 51%, positive predictive value of 14% and negative predictive value of 100%. It was also a significant predictor of haemodialysis (P < 0.05). All 13 patients who suffered renal damage and need for haemodialysis had arterial bicarbonate levels less than 22 mmol/L. The sensitivity of arterial bicarbonate as a predictor of haemodialysis is 100%, specificity 52%, positive predictive value 18% and negative predictive value 100%. Arterial bicarbonate was a significant predictor of duration of hospital stay (P < 0.05). 8/9 (88.8%) participants who were admitted for more than 2 weeks had arterial bicarbonate levels <22 mmol/L.

Arterial bicarbonate was a significant predictor of duration of ICU stay (P < 0.05). 116/140 (82.85%) participants required ICU stay. 45/58 participants who required ICU stay of more than 4 days had arterial bicarbonate levels <22 mmol/L.

Arterial base deficit as a predictor of mortality showed a P < 0.05 suggesting that base deficit is a significant predictor of mortality. Arterial base deficit values <–4 mmol/L were significantly associated with mortality. All 5 patients who expired had base deficit <–4 mmol/L. The sensitivity of arterial base deficit as a predictor of mortality associated with AP was 100%. The specificity of arterial base deficit as a predictor of mortality was 81%. The positive predictive value was 16% and the negative predictive value was 100%. Arterial base deficit was a significant predictor of need for haemodynamic supports (P < 0.05). 13/140 (9.28%) patients required haemodynamic support. 11/13 participants who required haemodynamic support had arterial base deficit < –4 mmol/L. The sensitivity of arterial base deficit as a predictor of need for haemodynamic supports was 85%, specificity 84%, positive predictive value 35% and negative predictive value 98%. It was a significant predictor of blood products transfused (P < 0.05). 13/19 (68.42%) participants who required blood product transfusion had arterial base deficit < –4 mmol/L. The sensitivity of arterial base deficit as predictor of blood product transfusion was 68%, specificity 85%, positive predictive value 42% and negative predictive value 94%. It was also a significant predictor of need for TPN (P < 0.05). 11/19 (57.89%) participants who required TPN had arterial base deficit < –4 mmol/L. The sensitivity of arterial base deficit in predicting the need for TPN was 58%, specificity 83%, positive predictive value 35% and negative predictive value 93%. It was also a significant predictor of need for ventilator support (P < 0.05). All 10 participants who required ventilatory support had arterial base deficit < –4 mmol/L. The sensitivity of arterial base deficit as a predictor of need for ventilator support was 100%, specificity 84%, positive predictive value 32% and negative predictive value 100%. It was a significant predictor of haemodialysis (P < 0.05). All 13 participants who required haemodialysis had arterial base deficit < –4 mmol/L. The sensitivity of arterial base deficit as a predictor of haemodialysis was 100%, specificity 86%, positive predictive value 42% and negative predictive value 100%. It was not a significant predictor of duration of hospital stay but was a significant predictor of duration of ICU stay (P < 0.05). 20/58 (34.4%) participants who required ICU stay for more than 4 days had arterial base deficit < –4 mmol/L. The sensitivity of arterial base deficit as a predictor of ICU stay was 34%, specificity 86%, positive predictive value 71% and negative predictive value 57%.

 Discussion



Predicting the severity and outcome of AP remains a challenge. Although multiple scoring systems are used to predict severe outcomes in AP, there is no consensus regarding the use of one or the other in clinical practice. Individual parameters of arterial blood gas have not been evaluated as predictors of severity and adverse outcome in AP. However, individual parameters of arterial blood gas form components of multiple scoring systems such as Ranson's and APACHE.[1],[2] One such report criticises arterial blood gas as an unnecessary tool in AP. However, this report, focused primarily on the blood gas and oxygenation-related parameters.[3] In majority of critical illnesses, there is ample evidence to implicate arterial blood gas parameters as important predictors of adverse outcome. It is seen that patients with organ dysfunction at the time of admission have higher mortality rates than those without.[4] The development of organ failure implies a poor prognosis. The highest mortalities are among patients with multiple and persistent organ failure.[5],[6]

In this study, 10 participants required ventilatory support and all 10 of them had pH value <7.35. All 19 participants who required blood product transfusions had pH value <7.35. All 13 participants who required haemodialysis had Arterial pH <7.35. 12/13 (92.3%) patients who were on haemodynamic support had bicarbonate levels <22 mmol/L. 17/19 (89.4%) participants who required blood product transfusion had bicarbonate level <22 mmol/L. All 10 participants who required respiratory support in the form of mechanical ventilation had bicarbonate level <22 mmol/L. All 13 patients who suffered renal damage and need for haemodialysis had bicarbonate levels than 22 mmol/L. 8/9 (88.8%) participants who were admitted for more than 2 weeks had bicarbonate levels <22 mmol/L. 45/58 (77.5%) participants who required ICU stay of more than 4 days had bicarbonate levels <22 mmol/L. 11/13 (84.6%) participants who required haemodynamic supports had base deficit < –4 mmol/L. 13/19 (68.42%) participants who required blood product transfusion had base deficit < –4 mmol/L. All 10 participants who required ventilatory support had base deficit < –4 mmol/L. All 13 participants who required haemodialysis had base deficit < –4 mmol/L. 20/58 (34.4%) participants who required ICU stay for more than 4 days had base deficit < –4 mmol/L. The above figures show that arterial blood gas parameters at admission are reliable markers of morbidity in the setting of AP. All 5 participants who expired had arterial pH < 7.35, bicarbonate levels <22 mmol/L and base deficit < –4 mmol/L. The association between acidosis and increased multiple organ failure in intensive care patients is long known.

Studies conducted in intensive care settings have documented the effects of metabolic acidosis on mortality. Metabolic acidosis (bicarbonate level of <18 mEq/L) was seen in 23/71 AP patients (32%) in a Brazilian study.[7] A study of 107 patients admitted to intensive care for various illnesses demonstrated a significantly severe metabolic acidosis in patients who expired.[8] Another study reported that severe metabolic acidosis occurs within the first 24 h in 6% of critically ill patients, and was associated with mortality rates in patients admitted to ICU.[9] In another prospective, observational study of 530 patients, pH at presentation could predict not only long-term mortality but also ICU admission, in-hospital mortality and mortality after 30-day follow-up.[10] Patients with a pH level of <7.39 showed a mortality rate of 37% after 12 months.

This study is important because a detailed evaluation of arterial blood gas analysis in AP has not been widely reported previously. The results suggest that pH, low bicarbonate and higher base deficit at presentation predicts an adverse outcome and worse prognosis in patients with AP, including the morbidity, long hospital and ICU stay, occurrence of organ failure and mortality. Therefore, carrying out an arterial blood gas analysis in patients with AP not only helps in the management of patients but may predict the adverse outcomes.

 Conclusion



In patients with AP, low arterial pH and bicarbonate levels and higher base deficit at presentation predict an adverse outcome with a higher frequency of organ failure and mortality. Thus, a simple diagnostic analysis of arterial blood gas in patients with AP can help predict an adverse outcome.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

References

1Ranson JH, Rifkind KM, Roses DF, Fink SD, Eng K, Spencer FC. Prognostic signs and the role of operative management in acute pancreatitis. Surg Gynecol Obstet 1974;139:69-81.
2Wu BU, Johannes RS, Sun X, Tabak Y, Conwell DL, Banks PA. The early prediction of mortality in acute pancreatitis: A large population-based study. Gut 2008;57:1698-703.
3Ward ST, Gilbert S, Mulchandani M, Garrett WV. Unnecessary investigations in patients with acute pancreatitis: Arterial blood sampling and serum biochemistry. Surgeon 2008;6:282-7.
4Johnson CD, Abu-Hilal M. Persistent organ failure during the first week as a marker of fatal outcome in acute pancreatitis. Gut 2004;53:1340-4.
5Buter A, Imrie CW, Carter CR, Evans S, McKay CJ. Dynamic nature of early organ dysfunction determines outcome in acute pancreatitis. Br J Surg 2002;89:298-302.
6Perez A, Whang EE, Brooks DC, Moore FD Jr, Hughes MD, Sica GT, et al. Is severity of necrotizing pancreatitis increased in extended necrosis and infected necrosis? Pancreas 2002;25:229-33.
7De Campos T, Braga CF, Kuryura L, Hebara D, Assef JC, Rasslan S. Changes in the management of patients with severe acute pancreatitis. Arq Gastroenterol 2008;45:181-5.
8Maciel AT, Park M. Differences in acid-base behavior between intensive care unit survivors and nonsurvivors using both a physicochemical and a standard base excess approach: A prospective, observational study. J Crit Care 2009;24:477-83.
9Jung B, Rimmele T, Le Goff C, Chanques G, Corne P, Jonquet O, et al. Severe metabolic or mixed acidemia on intensive care unit admission: Incidence, prognosis and administration of buffer therapy. A prospective, multiple-center study. Crit Care 2011;15:R238.
10Burri E, Potocki M, Drexler B, Schuetz P, Mebazaa A, Ahlfeld U, et al. Value of arterial blood gas analysis in patients with acute dyspnea: An observational study. Crit Care 2011;15:R145.