Article Text

Download PDFPDF

Chasing the base deficit: hyperchloraemic acidosis following 0.9% saline fluid resuscitation
  1. S Skellett,
  2. A Mayer,
  3. A Durward,
  4. S M Tibby,
  5. I A Murdoch
  1. Paediatric Intensive Care Unit, Guy's and St Thomas' Hospital, London SE1 9RT, UK
  1. Dr Durwardadurward{at}


Base deficit is a parameter often used to guide further treatment in acidotic children and is taken as a measure of how “sick” they are. Five children with septic shock are presented who had persisting base deficit after large volume resuscitation with 0.9% saline. Stewart's strong ion theory of acid–base balance is able to quantify the causes of metabolic acidosis and is used to show that our patients had a hyperchloraemic metabolic acidosis. We show how the chloride content of the saline loads given to our patients caused this hyperchloraemia. It is concluded that 0.9% saline and other chloride rich fluids may not be ideal resuscitation fluids; if used, clinicians must be aware of their potential to cause a persistent base deficit.

  • base deficit
  • 0.9% saline
  • resuscitation
  • hyperchloraemic metabolic acidosis

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Base deficit (BD) is traditionally used as a marker for metabolic acidosis and as such has gained a wide variety of clinical uses including prognostication,1 ,2 assessment of significant blunt trauma, shock, and regional hypoperfusion.3-5

It is therefore reasonable to assume that appropriate fluid resuscitation aimed at improving metabolic “well being”, by restoring tissue oxygenation and perfusion, should decrease the BD. As the magnitude of the BD may be correlated with mortality, an important yet overlooked issue concerning the crystalloid–colloid controversy is that the type of fluid used for resuscitation may influence acid–base status directly. Chloride rich solutions, such as 0.9% saline, used in large volumes can potentiate metabolic acidosis regardless of the underlying disease process.6-10 The mechanism for this chloride driven metabolic acidosis is easily explained by the strong ion theory of acid–base proposed by Stewart.11

In this report we briefly outline Stewart's theory and present clinical examples from five patients admitted to our paediatric intensive care unit, all with an appreciable base deficit due, in part, to large volume resuscitation (more than 40 ml/kg) with 0.9% saline.

The strong ion approach to acid–base balance

In contrast to the conventional Henderson–Hasselbalch approach, Stewart's theory states that three independent variables determine pH in plasma by changing the degree of water dissociation into hydrogen ions.11 These three variables are the strong ion difference (SID), the pCO2, and the charge from weak acids (ATOT). For example, a decrease in the SID, an increase in the pCO2 or ATOT all have an acidifying effect on plasma.

The effect that plasma chloride has on pH can be assessed by analysing the SID, which is calculated as the charge difference between the sum of measured strong cations (Na+, K+, Ca2+, and Mg2+) and measured strong anions (Cl, lactate). A strong ion is defined as one that is almost completely dissociated at physiological pH. As both Na+ and Cl are the major strong ions in plasma their ratio relative to one another largely determines the SID.

As shown in fig 1, an increase in the plasma Cl relative to Na+ decreases the plasma SID (normal values 38–42 mmol/l), thereby increasing the dissociation of water into hydrogen ions. In other words, the smaller the SID, the lower the pH.

Figure 1

Simplified diagram showing SID in a patient with normal plasma Na+ and Cl concentations (A) and in a patient with a hyperchloraemic metabolic acidosis (B).

Using the above principles, normal saline 0.9% has equimolar concentrations of Na+ and Cl (153 mmol/l) and therefore has an SID of 0. The administration of large quantities of normal saline will progressively lower the plasma SID, producing a hyperchloraemic metabolic acidosis. A solution of Ringer's lactate, which has an SID of 28 mmol/l, would decrease the pH to a lesser extent.

As the strong ion approach is based on the laws of mass conservation and electroneutrality, any discrepancy in the calculated SID from that directly measured allows for the detection of unmeasured anions in the form of a strong ion gap (SIG).6 ,7 ,12 ,13 This quantitatively represents the concentration of unmeasured strong anions as a result of tissue metabolic acid production. The strong ion approach has been validated experimentally and clinically in both healthy and critically ill patients.6 ,14 ,15

Clinical examples

To illustrate, we present five septic patients who were admitted to our unit, having been retrieved from their local hospitals following initial resuscitation. All presented with the clinical picture of septic shock16 and had been resuscitated with large volumes of normal saline. Blood samples were collected on admission from an indwelling arterial catheter and analysed for acid–base parameters (including lactate), electrolytes, and liver function tests; from these it was possible to calculate SID and SIG. Blood was also sent for microbiological culture and analysis for meningococcal PCR. Table 1 presents demographic data, including the Paediatric Index of Mortality and Glasgow Meningococcal Septicaemia Prognosis Score (GMSPS). Table 2 presents acid–base parameters taken on admission.

Table 1

Patient demographic data, outcome, and haematological indices on admission in five patients with septic shock

Table 2

Patient admission acid–base, blood lactate measurements, strong ion difference (SID), and strong ion gap (SIG) calculations


In order to determine whether the chloride gain from normal saline administration was sufficient to explain the post-resuscitation hyperchloraemia observed in the five patients (table 3), we performed the following calculations, assuming that plasma Clprior to normal saline treatment was normal and that chloride distributes throughout total body water:

Total body chloride = total body water × plasma Cl concentration
Chloride load from normal saline = volume normal saline given (litres) × 153 mmol/l (concentration of Cl in 0.9% saline)
Calculated Cl concentration (Clcalc) following normal saline load:
Embedded Image
Table 3

Patient blood chemistry measurements and calculated variables (Cl load and Clcalc)

From table 3, the Clcalc closely approximated the measured plasma Cl following fluid resuscitation in all but one patient. This would imply that the relative hyperchloraemia could be largely explained by the chloride gain from normal saline administration. The measured Cl in patient 4, who received over twice as much volume per kg as the other patients, exceeded the Clcalc by 7 mmol/l, indicating that either the plasma Cl was high prior to fluid resuscitation, or perhaps other mechanisms regarding chloride distribution are present following extremely large Cl loads.


Following the initial fluid resuscitation of the critically ill patient, clinicians are often faced with a grumbling, unexplained BD despite correction of hypoxia or hypovolaemia. In this situation it is tempting to chase the BD with further fluid boluses in order to improve metabolic “well being”. The aetiology of the acidaemia cannot be further explored using the traditional Henderson–Hasselbalch approach.

Using Stewart's approach on our patient group, we could show that the BD was a result of hyperchloraemia alone (decreased SID in the absence of significant lactate or unmeasured anion (SIG) concentrations). This relative hyperchloraemia could be accounted for by the large chloride load secondary to the volume resuscitation with normal saline.

Although restoration of intravascular volume remains a crucial and necessary goal of fluid resuscitation, failure to recognise the contribution of hyperchloraemia could lead to the BD becoming unreliable as a marker for effective resuscitation when large volumes of normal saline are used.

This then raises an interesting angle to the debate concerning the ideal resuscitation fluid. Neither normal saline nor colloid preparations such as 4.5% human albumin solution, which continue to be used,17 are physiological in the sense that both have an acidifying effect on the plasma. According to the strong ion theory, albumin, being a weak acid, increases the ATOT, therefore lowering pH.14 In addition, the electrolyte solution accompanying the 4.5% albumin preparation has similar Na+and Cl concentrations (100–160 mmol/l), which will further acidify plasma by reducing SID. Solutions with multicarbon anions (Ringer's lactate, Hartmann's, and Plasmalyte), which contain Cl and Na+ in concentrations similar to plasma, are more physiological and may be less likely to acidify plasma.

Although a persisting base deficit has been associated with increased mortality,12 ,18 to what degree a chloride driven acidosis influences mortality remains an open question. With this in mind, all one can conclude at present is that the use of large volumes of “non-physiological” chloride rich solutions such as normal saline or albumin, may potentiate metabolic acidosis, making BD interpretation misleading. Clinicians should be aware of the concept of a chloride driven acidosis and when faced with a persisting base deficit, once hypotension or hypoxia has been corrected, think twice before prescribing yet another fluid bolus of normal saline or 4.5% albumin solution to “chase” the base deficit.