Rapid Responses to:

EDITORIALS: Ketan K Dhatariya Diabetic ketoacidosis BMJ 2007; 334: 1284-1285 [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Give them Hartmann's

James S Dawson   (23 June 2007)

Diabetic Ketoacidosis

Hywel S Roberts   (24 June 2007)

Replacement of insulin, not of large volumes of fluid, is the primary treatment in diabetic ketoacidosis

Ernst Chantelau   (26 June 2007)

Diabetic Ketoacidosis

Robert Matz   (26 June 2007)

'Abnormal' saline is not superior to Hartmann's

Duncan L Wyncoll   (27 June 2007)

The pH of resuscitation fluids is important

Shondipon K Laha, Andrew J. Gosling   (28 June 2007)

Diabetic Ketoacidosis: A Major Problem

Simon Browne, Christopher Black, Sujoy Ghosh, Iqbal Malik, Andrew Collier   (28 June 2007)

Hartmann’s Solution and 0.9% Saline are Both Unsuitable in Severe DKA

Roger Townsend   (28 June 2007)

A consensus with a curious recommendation?

Craig G Morris, James Low   (1 July 2007)

Diabetic ketoacidosis: are tradition and physiology enough to discount the need for evidence?

Michael C. Reade   (14 July 2007)

Fluid Resuscitation in Diabetic Ketoacidosis

Kevin J Hardy, Richard Griffiths   (16 July 2007)

Re:The pH of resuscitation fluids is important

David J Miller   (23 July 2007)

Would Normal Saline be licensed today?

Nicholas Levy, Dr Stuart Lowe   (11 October 2007)

Maintenance versus Resuscitation Fluids

Gareth Hynes   (14 April 2008)

Give them Hartmann's 23 June 2007
James S Dawson, SHO Anaesthesia and Critical Care Birmingham Send response to journal: Re: Give them Hartmann's	Discussions on intravenous fluids in various situations have been ongoing for countless years, and will probably continue for countless more. For this reason alone, I’m surprised Dhatariya makes an error when describing the millimolar concentrations of sodium and chloride in 0.9% saline – 154 mmol/L for each, not 150 mmol/L. As for the list of ‘contraindications’ for using Hartmann’s solution in patients with diabetic ketoacidosis, and the apparent overwhelming support for 0.9% saline, I am sure I will not be the only reader who finds this editorial unbalanced. 1. High lactate to pyruvate ratio is presumably related to poor tissue perfusion (dehydration) in this group of patients, and the study quoted looked at a very mixed group of patients, and found that a persistently high lactate to pyruvate ratio is associated with a worse outcome. But as we know, patients with DKA rapidly clear their lactate once their circulatory failure is treated – whether with Hartmann’s or saline. 2. Hartmann’s solution contains lactate as a buffer, so it will clearly elevate serum lactate. Hyperlactaemia in patients with DKA is common, for reasons described above. The metabolic acidaemia seen in patients with DKA results largely from ketones. Hyperlactaemia per se is not a problem, as most patients with DKA have functioning livers and can rapidly metabolise lactate; patients in renal failure on the ITU are most often given renal replacement therapy using lactate-buffered solutions. Lactate can be converted into glucose, via the Cori cycle, which should not pose any problems, since the other arm of treatment in DKA should include insulin. As the rate of fluid replacement falls over the course of treatment, the contribution Hartmann’s solution makes to plasma glucose will be minimal. 3. Hyperkalaemia is seldom apparent on presentation, despite what the text-books say, and we all know these patients are massively depleted of potassium. As soon as insulin therapy is commenced aggressive potassium replacement is mandatory. As intravenous access should be secured and fluids commenced prior to insulin therapy, it may be wise to use a fluid such as 0.9% saline for the first litre, but then switching to Hartmann’s and giving a potassium infusion would be my ITU management. 4. Bicarbonate does indeed worsen intra-cellular acidosis, but the quantities of bicarbonate in 8.4% solutions compared to that which is liberated from the lactate in Hartmann’s solution are an order of a magnitude or two different. 5. Hyponatraemia, like hypernatraemia, is avoidable by careful, frequent monitoring, and is more likely to result from glucose (dextrose) solutions than Hartmann’s. On the intensive care, we spend a lot of time giving water to help the kidneys clear massive amounts of sodium which has been administered by well-meaning doctors. I do agree with the authors ideal about comparing 0.9% saline with Hartmann’s in a randomized control trial in this group of patients, and I am convinced Hartmann’s would come out far superior, and become the fluid of choice in this situation, and in many others. Competing interests: None declared
Diabetic Ketoacidosis 24 June 2007
Hywel S Roberts, Specialist Registrar in Intensive Care Medicine University Hospital of Wales, Cardiff, UK, CF14 4XW Send response to journal: Re: Diabetic Ketoacidosis	As a senior trainee approaching the end of my training in Intensive Care Medicine, who has also spent 10 months as a trainee in Diabetes and Endocrinology, I was interested to read Dr Dhatariya’s editorial (1) on the differences in the management of Diabetic Ketoacidosis by Intensive Care Physicians and Diabetologists. Whilst I agree with Dr Dhatariya that there are sound physiological reasons why Normal Saline should be used as the initial fluid in the resuscitation of diabetic ketoacidosis I must take issue with several of the matters discussed. Firstly, I do not accept that the “preferred” fluid for Intensive Care Physicians is “commonly Hartmann’s solution". I do not know of any Intensive Care Physician that uses Hartmann’s solution in such a way and I suspect that Dr Dhatariya’s experiences may be related to Critical Care provision by some physicians largely specialising in Anaesthetics. I suspect that most modern Critical Care specialists have at least as good an understanding of the issues discussed as a typical Diabetologist, along with a great deal of practical experience with the management of complex fluid and electrolyte disorders. As a specialist group, I would imagine that there has not been a need for critical care physicians to write their “own” guidelines for management of diabetic ketoacidosis as most would be familiar with the guidelines published by other specialities. Secondly, I would take great issue with the statement that “..if a high chloride concentration is found then the base deficit can be safely ignored”. Strong Ion Difference theory (imperfect as it is and difficult to apply in practice) suggests that it is the difference between the Sodium and Chloride concentrations which is responsible for the metabolic acidosis (2) . As such it is difficult to quantify exactly how much of a given base deficit can safely be attributed to Hyperchloraemia. It is important to keep an open mind regarding the cause of a significant base deficit if potentially serious causes are not to be missed. Acidosis due to acute kidney injury or lactic acidosis from sepsis are relatively common in this patient population and should be treated aggressively. Finally, whilst I agree with Dr Dhatariya that ideally a randomised study would be performed to assess the relative safety of Hartmanns and Normal Saline, experience from other large trials comparing resuscitation fluids (3), along with the relatively low mortality of appropriately treated diabetic ketoacidosis would suggest that an extremely large number of patients would be required to show any significant difference. The resources required would be better directed towards researching aspects of the disease that still cause mortality and are poorly understood – for example the causes of cerebral oedema. 1) Dhatariya KK. Diabetic Ketoacidosis. BMJ 2007;334:1284-5 2) Fink MP, Abraham E, Vincent JL, Kochanek PM. Acid Base Disorders in Textbook of Critical Care 5th Edition. Pub: Elsevier Saunders, 2004 3) The SAFE study. N Engl J Med 2004;350:2247-56 Competing interests: None declared
Replacement of insulin, not of large volumes of fluid, is the primary treatment in diabetic ketoacidosis 26 June 2007
Ernst Chantelau, Professor of internal medicine Department od Endicrinology, Diabetes and Rheumatology, University of Düsseldorf/Germany Send response to journal: Re: Replacement of insulin, not of large volumes of fluid, is the primary treatment in diabetic ketoacidosis	Doctor Dhatariya's contention that "the primary treatment in diabetic ketoacidosis is replacement with large volumes of fluid" is unfounded. Moreover, it is grossly misleading, since recent research has shown that large volumes of fluid increase the risk of brain edema(1-4), the leading cause of death from diabetic ketoacidosis. Sincerely, Ernst Chantelau, MD, PhD 1)Edge JA, Jakes RW, Roy Y. et al. The UK case-control study of cerebral oedema complicating diabetic ketoacidosis in children. Diabetologia 2006;49:2002-2009 2)Rosenbloom A. Cerebral edema in diabetic ketoacidosis. J Clin Endocrinol Metab 2000;85:507-508 3)Finberg L. Appropriate therapy can prevent cerebral swelling in diabetic ketoacidosis. J Clin Endocrinol Metab 2000;85:508-509 4)Muri A. Cerebral edema in diabetic ketoacidosis: a look beyond rehydration. J Clin Endocrinol Metab 2000;85:509-513 Competing interests: None declared
Diabetic Ketoacidosis 26 June 2007
Robert Matz, Professor--Medicine Mount Sinai School of Medicine10029 Send response to journal: Re: Diabetic Ketoacidosis	Both Dawson and Roberts Rapid Response to Dhatariya's editorial are correct. "Normal saline"[0.95 NaCl] is about as unphysiologic a solution as one can imagine.It contains excess chloride,as compared to serum by about 45-50 mEq/l.In DKA where chloride has been retained by the kidney to accompany reabsorbed Na+ and K+ while the weak organic anions,the ketone bodies, have been preferentially lost in the urine accompanying H+ as well as Na+ and K+ this results in a "masked hyperchloremic"acidosis on presentation-i.e.the acidosis is not fully accounted for by the anion gap on presentation[2,3].This is unmasked when insulin begins to allow conversion of the ketones to bicarbonate.This,already present hyperchloremia is exacerbated by the excess chloride infused with 0.9% saline.While the normal kidney will ultimately correct this imbalance there is no need for it to occur. Almost all patients with DKA have total body K+ depletion regardless of the initial K+ value and the addition of 5 mEq of K+ /L in therapy will not precipitate worsening hyperkalemia. There is much speculation about the effect of hypotonic fluids in treating DKA-especially in the pediatric population,but very little light.As Krane,et.al [4] demonstrated with CT scan cerebral edema is present at the outset of therapy in many children with DKA.These children with DKA are further only rarely hyperosmolar.In the adult population with hyperglycemic hyperosmolarity,where one might expect to see cerebral edema as the are treated,confirmed cases are about as frequent as confirmed sightings of the Loh Ness monster[5]. Dawson correctly noted that the amount of bicarbonate generated from the administered lactate is an order of magnitude less than when bicarbonate is administered,and does not cause the abrupt shifts in pH seen with bicarbonate. If one is unhappy with a lactate containing solution ten there are commercially available solutions[at least in th U.S.] such as Plasmalyte 148 which provide Acetate and gluconate as the bicarbonate precurso instead of lactate which are equally as effective aslong as the liver, and perfusion are intact and insulin is available.These solutions contain 140mEq/l ofNa+,98 mEq/l ofCl-,and 5 mEq/l of K+. It is time to put 0.9% NaCl to rest. Ref. 1.Dhatariya KK.Diabetic ketoacidosis.BMJ 2007;334;1284-1285. 2.Androgue HJ, Wilson H Boyd AE,Suki WN,Eknoyan G.Plasma acid-base patterns in diabetic ketoacidosis.N Engl J Med1982;307;1603-1610. 3.Androgue HJ,Eknoyan G,Suki WK.Diabetic ketoacidosis:role of the kidney inacid-base homeostasisre-evaluated.Kidney Int.1984;25;591-598. 4.Krane EJ, Rackoff MA,Wallman JA,et.al..Subdural brain swelling during treatment of diabetic ketoacidosis.N Engl J Med.1985;312;1147. 5. Matz R.Hyperglycemic hyperosmolar syndrome.In.Porte D,Sherwin RS,Baron A,eds.Ellenberg and Rifkinn's Diabetes Mellitus.6th Ed.McGraw Hill.2003;587-599 Competing interests: None declared
'Abnormal' saline is not superior to Hartmann's 27 June 2007
Duncan L Wyncoll, Consultant Intensivist Guy's & St Thomas' NHS Foundation Trust Send response to journal: Re: 'Abnormal' saline is not superior to Hartmann's	I was surprised to read the editorial from Dhatariya (1) where 0.9% saline was pushed so strongly as the optimal choice of fluid resuscitation in diabetic ketoacidosis (DKA). Many of the reasons given in promotion of 0.9% saline are just not clinically relevant in most patients with DKA. For example, the small amount of potassium in Hartmann's solution is still not enough for most patients, and after the immediate resuscitative phase hypokalaemia is a predominant problem. Also the issues raised around cerebral oedema, may be much more relevant to children and not so typical in adults. It is important that we are realistic, 0.9% saline is not normal, but very abnormal and not remotely physiological. It inevitably causes hyperchloraemic metabolic acidosis, and it is incorrect to say that it is mild, transient and not associated with adverse outcomes. In a number of different situations 'abnormal saline' (0.9%) has been shown to be inferior to physiologically balanced solutions. (2-5) Whilst I accept that 0.9% saline may be the currently recommended solution to use in the immediate resuscitative phase for fluid replacement in DKA - i.e. the first few hours or the first 2-3 litres for an adult. Surely the most physiologically sound action is then to switch to a balanced electrolyte solution to avoid iatrogenic acidosis. This is why I cannot agree with Dhatariya that a randomised trial of 0.9% saline versus Hartmann's solution is a rational concept - use the right fluid, at the right time, to achieve the right goal. Hartmann's is far from a perfect solution, since it is hypotonic, and the lactate cannot be metabolised by some patients. Consequently, opportunities should always be taken to encourage industry to manufacture a better normotonic, balanced, physiological solution - something which is not currently available in the UK unless 'dialysis type' fluids are used off-license. References 1. Dhatariya KK. Diabetic Ketoacidosis. BMJ 2007; 334: 1284-5. 2. O’Malley et al. A randomized, double-blind comparison of lactated Ringer's solution and 0.9% NaCl during renal transplantation. Anesth Analg 2005; 100: 1518-24. 3. McFarlane C, Lee A. A comparison of Plasmalyte 148 and 0.9% saline for intra-operative fluid replacement. Anaesthesia 1994; 49: 779-81. 4. Waters JH, et al. Normal saline versus lactated Ringer's solution for intraoperative fluid management in patients undergoing abdominal aortic aneurysm repair: an outcome study. Anesth Analg 2001; 93: 817-22. 5. Williams EL, Hildebrand KL, McCormick SA, Bedel MJ. The effect of intravenous lactated Ringer's solution versus 0.9% sodium chloride solution on serum osmolality in human volunteers. Anesth Analg 1999; 88: 999-1003. Competing interests: None declared
The pH of resuscitation fluids is important 28 June 2007
Shondipon K Laha, Consultant in Anaesthesia and Critical Care Medicine Lancashire Teaching Hospitals NHS Foundation Trust, Andrew J. Gosling Send response to journal: Re: The pH of resuscitation fluids is important	We were interested to read Dr Dhatariya's comments (1) on the value of 0.9% saline as opposed to the use of compund sodium lactate (Hartmann's solution). We would like to dispute the supposition that he makes that large volumes of 0.9% saline induce a mild acidosis. The term pH reflects the equation -log10 [H+]. Hence a change of the pH of 1 is a tenfold change in hydrogen ions. Normal pH of 7.4 approximates 40 nanomoles of H+ ions. Hartmann's solution has a ph of 6 or 1000 nanomoles of H+ ions while 0.9% saline has a pH of 4.5 or 31 600 nanomoles. This is 30 times the H+ concentration of Hartmann's and almost a 1000 times normal blood. We fail to see how this can be ignored in patients who are already acidaemic and are then given several litres of saline. References 1.Dhatariya K..K. Diabetic Ketoacidosis. BMJ 2007; 334: 1284-1285 Competing interests: None declared
Diabetic Ketoacidosis: A Major Problem 28 June 2007
Simon Browne, Senior House Officer The Ayr Hospital, Christopher Black, Sujoy Ghosh, Iqbal Malik, Andrew Collier Send response to journal: Re: Diabetic Ketoacidosis: A Major Problem	Diabetic Ketoacidosis: A Major Problem Dr. Simon Browne, Dr.Christopher Black, Dr. Sujoy Ghosh, Dr.Iqbal Malik, Dr. Andrew Colier Diabetes Day Centre: The Ayr Hospital, Ayr, Ayrshire, Scotland Correspondence E-mail: drsujoyghosh@rediffmail.com We have read with interest the article on diabetic ketoacidosis.[1] We did an audit on patients who presented with diabetic ketoacidosis (DKA) between January 2005 and January 2006. We had a gut feeling that admissions with DKA was on the rise. We found that 36 patients (amounting to 40 admissions) were admitted with DKA during the aforementioned period. 32 had type 1 diabetes and only 4 had type 2 diabetes. Most patients presented with moderate to severe degree of DKA. Not surprisingly most patients had a poor glycaemic control, as reflected by high HbA1c. (about 50% had a HbA1c>10). We found that the cause for DKA was multifactorial. However vomiting, infection, alcohol excess and non compliance to insulin regimen were the common precipitating causes. There were at times irregularities in the implementation of the DKA protocol. Unfortunately 2 patients died (1 unrelated to DKA). Patients received education (pre-discharge) with regards to management of diabetes and prevention of any further episode of DKA. We were surprised to find that most of the patients who presented with DKA had had diabetes for more than 5 years and had not had DKA in the past. We believe that the number of patients admitted with DKA has suddenly increased in recent times, inspite of the fact that patients are receiving intensive education with regards to management of diabetes. We believe that the level of care provided (both primary care and hospital based diabetes clinics) has remained the same (if not improved) for type 2 diabetes but we believe that type 1 diabetes education has taken a back seat in recent times. In addition we believe that patients tend to forget what they are taught and therefore we need to re-educate our patients (especially sick day rules). What could then account for the sudden rise in the incidence of DKA? Could it be that infection (particularly viral) is more prolonged and severe in recent times or could it be due to the use of newer insulin analogues? Could it be possible that the analogues are less effective (under stressful conditions) in inhibiting lipolysis? Further studies are required to answer the question. [1] Dhatariya K K. Diabetic ketoacidosis. BMJ. 2007 Jun 23;334(7607):1284-5. Competing interests: None declared
Hartmann’s Solution and 0.9% Saline are Both Unsuitable in Severe DKA 28 June 2007
Roger Townsend, Consultant Intensivist University Hospitals Coventry & Warwickshire, Clifford Bridge Road, Coventry, CV2 2DX Send response to journal: Re: Hartmann’s Solution and 0.9% Saline are Both Unsuitable in Severe DKA	Dhatariya’s editorial [1], arguing that 0.9% Saline (NS) should be used in preference to Hartmann’s solution in Diabetic Ketoacidosis (DKA) is timely. However, his argument against using Hartmann’s solution should be balanced against the good reasons for avoiding excessive saline administration in critically ill patients, namely sodium overload and oedema together with hyperchloraemic acidosis [2]. Recently there has been renewed interest in the Stewart’s approach to acid-base physiology and the phenomenon of hyperchloraemic acidosis [3]. In patients with severe DKA, who present with profound metabolic acidosis and hypovolaemia, resuscitation with large volumes NS often results in an equally profound hyperchloraemic acidosis. Dhatariya’s contention that acidosis caused by large volumes of NS is mild and transient and not associated with adverse outcomes must be challenged. Any metabolic acidosis is a powerful marker of poor prognosis in the critically ill [4]. Moreover, there is mounting evidence that hyperchloraemic acidosis may cause abdominal dysfunction [5], renal dysfunction [5] and, possibly, coagulation abnormalities [6]. In addition, there is evidence from animal models of haemorrhage[7] and sepsis [8] that NS is inferior compared to ‘balanced’ fluids (i.e. those containing more physiological concentrations of chloride) with regard to survival. Perhaps the greatest risk of using large volumes of NS in this severe DKA is that a persisting base deficit may be interpreted as evidence of ongoing ketosis or hypovolaemia, with a resultant inappropriate change in management, as has been reported previously in a different context [9]. Neither Hartmann’s solution nor NS are ideal as the sole replacement fluid in severe DKA. However, until we have an isotonic, physiological solution licensed for intravenous use we should use the fluids we have available in regimens based on the pathophysiology of DKA. It is inappropriate to assume that all patients with DKA have the same degree of fluid loss and hypovolaemia and, therefore, prescribe fluid regimens which are inflexible in respect of rate and volume. It is similarly inappropriate to advocate a single replacement fluid for all patients in this context. The composition of urine in DKA is approximately 70 mmol/L for both sodium and potassium [10]. Therefore, from a physiological perspective the ‘ideal’ replacement fluid would be half-‘normal’ saline (i.e. what the patient is losing). This would correct dehydration and avoid sodium overload and hyperchloraemia. However, it would not rapidly correct intravascular hypovolaemia and may cause a rapid fall in plasma osmolality and lead to cerebral oedema. In DKA, fluid is lost from all three body fluid compartments (intravascular, interstitial and intracellular) and the greatest threat to life is from intravascular fluid depletion. For a given volume of NS, only one-third remains intravascular, with the remaining two-thirds expanding the interstitial space (inevitably causing oedema). Moreover, NS administration does not address the intracellular dehydration of DKA. The most efficient fluid for correction of hypovolaemia is a colloid, as most of this is retained in the intravascular space. Unfortunately, the majority of colloids in UK practice are suspended in NS, risking the hyperchloraemia and sodium overload we wish to avoid. However, for a given degree of hypovolaemia, a smaller volume of colloid will be required for resuscitation, resulting in lesser amounts of sodium and chloride being administered. In severe DKA with circulatory shock, intravascular volume depletion should be corrected with a suitable volume of colloid, guided by conventional endpoints (heart rate, blood pressure and capillary refill), together with simultaneous administration of water (as 5% Dextrose) to replace losses from the interstitial and intracellular space [11]. The amount of glucose in 5% dextrose is small and will not significantly exacerbate the pre-existing hyperglycaemia. In less severe DKA, where patients may be managed outside of a high dependency environment, NS may be the most suitable fluid of those currently available. However, a randomised study to compare NS with Hartmann’s in this context would be inappropriate as neither fluid is ideal for managing severe DKA. 1. Dhatariya, K.K., Diabetic ketoacidosis. BMJ, 2007. 334(7607): p. 1284-5. 2. Gosling, P., Salt of the earth or a drop in the ocean? A pathophysiological approach to fluid resuscitation. Emerg Med J, 2003. 20(4): p. 306-15. 3. Story, D.A., Hyperchloraemic acidosis: another misnomer? Crit Care Resusc, 2004. 6(3): p. 188-92. 4. Gunnerson, K.J., et al., Lactate versus non-lactate metabolic acidosis: a retrospective outcome evaluation of critically ill patients. Crit Care, 2006. 10(1): p. R22. 5. Williams, E.L., et al., The effect of intravenous lactated Ringer's solution versus 0.9% sodium chloride solution on serum osmolality in human volunteers. Anesth Analg, 1999. 88(5): p. 999-1003. 6. Gan, T.J., et al., Hextend, a physiologically balanced plasma expander for large volume use in major surgery: a randomized phase III clinical trial. Hextend Study Group. Anesth Analg, 1999. 88(5): p. 992-8. 7. Healey, M.A., et al., Lactated ringer's is superior to normal saline in a model of massive hemorrhage and resuscitation. J Trauma, 1998. 45(5): p. 894-9. 8. Kellum, J.A., Fluid resuscitation and hyperchloremic acidosis in experimental sepsis: improved short-term survival and acid-base balance with Hextend compared with saline. Crit Care Med, 2002. 30(2): p. 300-5. 9. Parekh, N., Hyperchloremic acidosis. Anesth Analg, 2002. 95(6): p. 1821; author reply 1821-2. 10. Atchley D.W., Leob R.F., Richards D.W. Jr et al, On Diabetic Acidosis; A detailed study of the electrolyte balance following the withdrawal and re-establishment of insulin therapy. J Clin Invest, 1933. 12: p. 297-326. 11. Hillman, K., Resuscitation in diabetic ketoacidosis. Crit Care Med, 1983. 11: p. 53-54. Competing interests: None declared
A consensus with a curious recommendation? 1 July 2007
Craig G Morris, Consultant intensivist and anaesthetist Derby Hospitals, DE1 2QY, James Low Send response to journal: Re: A consensus with a curious recommendation?	We read with interest the recent editorial on the use of normal saline during the initial resuscitation stages of treating patients with severe life threatening diabetic keto acidosis. We would like to thank the authors for highlighting this common medical emergency and the complicated issues that are raised in terms of the fluid management of these patients in the early resuscitation period. We would however like to challenge some of the conclusions that have been reached in this editorial regarding the appropriateness of normal saline in these conditions. This editorial makes the valid point that we will not have a prospective, randomised, controlled trial to decide whether Hartmann’s or saline is safest or superior. However, to follow this up with a consensus statement, however well intentioned, as “case closed” is far from acceptable. We would strongly argue that Hartmann’s constitutes an excellent choice for volume loading in DKA and reject this editorial’s determinations. Clinicians who use Hartmann’s do so in recognition of its desirable electrolyte and osmolar composition. The racemic lactate contained within (29mmoll-1 ) is metabolised to glucose (60%) and oxidised (40%), both reactions consuming protons equivalent to approximately 50mmoll-1.1 While we agree the effect of acidic chloride anion accumulation on outcome is largely unresolved, although it may not be benign, it seems inconsistent that this is glossed over as transient and of no consequence, while accumulating lactate, from administered Hartmann’s solution, is seen as damaging. During renal replacement therapy for critically ill patients we typically administer solutions of 40mmoll-1 lactate concentration at volumes of 2- 5.0 litres per hour! The liver and kidneys normally metabolise 1500 mmol lactate per day and have been demonstrated to be able to metabolise up to 100mmolhr-1. 2 The elevated lactate: pyruvate levels are indicative of type A lactic acidosis rather than exogenous lactate administration, and it is a common misconception that exogenous lactate from Hartmann’s contributes to lactic acidosis: this is not cause and effect, a point made by the reference from the journal “Shock”.3 As will be familiar to a diabetologist this change in lactate: puruvate level is also mimicked by a change towards a betahydroxybutyrate: acetoacetate ratio, but again this generally indicates an ischaemic cell environment. The mantra that bicarbonate therapy worsens intracellular acidosis is far more complex than this editorial suggests and is dependent on local buffering capacity, anaemia and speed of administration.4, 5 This perennial bicarbonate debate will not be resolved in the correspondence section, but Hartmann’s with its gradual consumption of modest amounts of protons approaches a very physiological process and is a million miles away from rapid infusions or boluses of concentrated sodium bicarbonate! We can only echo the concerns expressed regarding cerebral oedema and it is clear that fluid and electrolyte administration is less than optimal for many acute medical patients as has been reported before in the BMJ. However, the effective osmolality of Hartmann’s (approximately 285 mOsmoll -1) is appropriate for clinical use, and the presence of hyponatraemia on admission tells a different story- after all, how does an osmotically dehydrated patient become hyponatraemic without sodium loss? Admission hyponatraemia is a good marker for natriuresis and an element of adrenal insufficiency, and coupled with non-osmotic AVP release, cerebral oedema becomes a marker for lead time bias and associated defects of water handling, not an indictment of Hartmann’s. In reality most DKA patients get saline loading until shock resolves then a dextrose or dextrose/ saline infusion is started as a source of glucose, in the presence of insulin, to clear ketones. If we want to produce osmolar shifts and cerebral oedema this is a great model for doing it and we would far rather see Hartmann’s or 0.9% saline with dextrose added to it. Finally, plasma osmolality is a misleading concept in DKA as an intracellular glycopenia reduces effective intracellular tonicity and cell volume. As insulin is administered, the osmotic (or tonicity) gradient across the cell tends to reduce but cell volume tends to increase. The key message here, with most questions of osmolality, is to do it slowly and monitor closely. In summary we can only agree with the authors that “The question of which fluid replacement is optimal…is, therefore, still unanswered”. We recognise the “potential dangers” of using Hartmann’s, as with any intervention including insulin itself, but we reject the basic science arguments as proposed and would suggest that Hartmann’s was and still is an excellent choice for volume resuscitation in DKA, and maintenance fluid until establishing oral intake. References 1 White SA, Goldhill DR. Is Hartmann’s the solution? Anaesthesia 1997;52:422- 7 2 Kierdorf HP, Leue C, Arns S. Lactate or bicarbonate-buffered solutions in continuous extracorporeal renal replacement therapies. Kidney Int 1999;S72:S32- 6 3 Suistomaa M, Ruokonen E, Kari A, Takala J. Time-pattern of lactate and lactate to pyruvate ratio in the first 24 hours of intensive care emergency admissions. Shock 2000;14:8-12 4 Levraut J, Giunti C, Ciebiera JP et al. Initial effect of sodium bicarbonate in intracellular pH depends on the extracellular nonbicarbonate buffering capacity. Crit Care Med 2001;29:1033- 9 5 Levraut J, Garcia P, Giunti C et al. The increase in CO2 production induced by NaHCO3 depends on blood albumin and hemoglobin concentrations. Intensive Care Med 2000;26:558- 64 Competing interests: None declared
Diabetic ketoacidosis: are tradition and physiology enough to discount the need for evidence? 14 July 2007
Michael C. Reade, Attending Intensivist & Visiting Instructor University of Pittsburgh Medical Center, 3550 Terrace St, Pittsburgh PA 15261, USA Send response to journal: Re: Diabetic ketoacidosis: are tradition and physiology enough to discount the need for evidence?	The BMJ’s publication of Dr Dhatariya’s editorial 1 is perplexing in the era of evidence-based medicine. There are undoubtedly a number of physiological arguments in favour of using 0.9% saline in the resuscitation of patients with diabetic ketoacidosis (DKA), but as adroitly pointed out by authors of the accompanying Rapid Responses, many have tenuous validity, and there are equally cogent arguments in favour of Hartmann’s solution or even more ‘physiological’ fluids. As Dr Dhatariya admits, ‘the question of which fluid replacement is optimal .. is .. unanswered’. What follows is essentially a one-sided scientific justification of tradition-based medicine. Fortunately, two independent traditions appear to have emerged: that of endocrinologists (who favour saline in their guidelines) and that of intensivists and emergency physicians (who often favour Hartmann’s in their practice). In many hospital systems most patients with DKA are stabilised by intensivists and emergency physicians, usually with better access to electrolyte and blood gas monitoring than is available to endocrinologists on a regular ward, implying that there must be at least some validity to the contrary argument. This sets ideal conditions for a clinical trial. The bottom line is that without a clinical trial we will never know which, if any, fluid is superior. As noted in the Rapid Responses, in a condition with such low treated mortality, demonstrating survival benefit may be difficult. However, valid surrogate outcomes should be easily assessed. Dr Dhatariya’s contention that equipoise for such a trial cannot exist (‘because of the potential dangers of Hartmann’s solution’), based on contentious physiological arguments alone and in the face of common alternate practice, harks back to a less scientific (and less safe) era of medicine. 1. Dhatariya KK. Diabetic ketoacidosis. BMJ 2007;334:1284-5 Competing interests: None declared
Fluid Resuscitation in Diabetic Ketoacidosis 16 July 2007
Kevin J Hardy, Consultant Diabetologist L35 5DR, Richard Griffiths Send response to journal: Re: Fluid Resuscitation in Diabetic Ketoacidosis	We read with interest Dhatariya’s editorial [1] arguing for use of 0.9% saline in preference to Hartmann’s solution in the management of diabetic ketoacidosis (DKA). Certainly, recent DKA guidelines from the American Diabetes Association [2] and the Association of British Clinical Diabetologists [3] and the Scottish national protocol for the management of diabetic ketoacidosis [4] advocate use of 0.9% saline. But is there ‘a large body of supporting evidence’ for this recommendation, as Dhatariya claims? Dhatariya argues for saline on the basis of ‘theoretical’ disadvantages with Hartmann’s solution: its glucose-generating lactate content, its potassium content, its bicarbonate content and its reduced sodium content and he dismisses the hyperchloraemic acidosis that saline generates as ‘mild and transient’. Using similar arguments, however, the excessive and oedema-generating sodium load of large volumes of normal saline, particularly when coupled with insulin’s anti-natiuretic effects might be said to favour Hartmann’s solution. And although the bicarbonate in Hartmann’s solution might theoretically be problematic, for already acidotic patients, is it not preferable to the 30-fold increased concentration of hydrogen ions in normal saline? It is difficult to evaluate the effects of hyperchloraemic acidosis in already severely acidotic patients, but it may not be as innocuous as Dhatariya implies [5]. In the context of intravenous insulin therapy, we would argue that glucose generation from lactate is irrelevant (indeed, glucose solutions are administered intravenously in the latter stages of DKA management in the above guidelines) and since patients are invariably total body potassium deplete in DKA and fluid replacement is guided by electrolyte results, used appropriately, the physiological concentrations of potassium in Hartmann’s solution are more likely to do good than harm. We conclude that the best replacement fluid in DKA is far from clear and that further research using clinically relevant outcomes should be undertaken to guide optimal management of this serious and not uncommon condition. 1. Dhatariya KK. Diabetic ketoacidosis. BMJ 2007;334:1284-5. 2. Kitabchi AE, Umpierrez GE, Murphy MB, Kreisberg RA. Hyperglycemic crises in adult patients with diabetes: a consensus statement from the American Diabetes Association. Diabetes Care 2006;29:2739-48. 3. Savage MW, Kilvert A. ABCD guidelines for the management of hyperglycaemic emergencies in adults. Pract Diab Int 2006;23:277-31. 4. McGeoch SC, Hutcheon SD, Vaughan SM, et al. Pract Diab Int 2007;24:257-61. 5. Williams EL, Hillebrand KL, McCormick SA, Bedel MJ. The effect of intravenous lactated Ringer’s solution versus 0.9% sodium chloride solution on serum osmolality in human volunteers. Aneth Analg 1999;88:999 -1003. Competing interests: None declared
Re:The pH of resuscitation fluids is important 23 July 2007
David J Miller, Hon Research Fellow Biomed Sci, Glasgow University G12 8QQ Send response to journal: Re: Re:The pH of resuscitation fluids is important	Many of my basic science colleagues will share my anxiety at aspects of this discussion. The pH of a saline per se is not really an issue. It is the combination of a solution's buffer capacity together with its pH that is important. The pH of an unbuffered saline may well be as low as 5 - sufficient CO2 is dissolved (from the air) and disocciated to cause this. However, the buffering power will be essentially trivial. But infusing a solution like Hartmann's that does contain many millimolars of a significant pH buffer represents a significant acid (or alkaline) load if the pH differs from that of blood. The extent of that load is indeed logarithmically affected by the pH difference. However, the homeostatic power of respiratory, excretory and metabolic mechanisms is huge ... depending on the physiological status of the recipient of the infusion, of course. Quite why the clinical infusates are not routinely a closer match to normal plasma values remains a bit of mystery to non-clinicians like me. Sydney Ringer and his successors sorted out the composition of the plasma as regards electrolytes. pH buffering has been well understood for decades by physiologists and chemists. Metabolic details and the special constraints imposed by certain disease conditions are surely understood by both physiologists and physicians. In extremis we could indeed sensibly and safely resort to making up a solution of 0.9% of table salt and distilled water. But nearly every clinical application will utilise commercial products that could surely be far more physiologically 'correct'. If not, why not? Competing interests: None declared
Would Normal Saline be licensed today? 11 October 2007
Nicholas Levy, Consultant in Anaesthesia and Critical Care West Suffolk Hospital, IP33 2QZ, Dr Stuart Lowe Send response to journal: Re: Would Normal Saline be licensed today?	Dear Sir We read with interest the recent editorial by Dhaetariya(1). In his editorial, the use of Hartmann’s solution is bought into question and most of the physiological arguments are eloquently rebuked in the ensuring rapid response correspondence. One of his most contentious comments is that there is unlikely to be a randomised study between normal saline and Hartmann’s solution due to the potential dangers of Hartmann’s solution. This comment is highly provocative as it suggests that those professionals who treat their diabetic patients with Hartmann’s solution are essentially being negligent, and as the ensuring correspondence shows there are a number of intensivists who choose to use Hartmann’s as part of the fluid regime to resuscitate patients with diabetic ketoacidosis (DKA). Dr Dhaetariya questions the validity of use of Hartmann’s solution, which whilst we agree is not the perfect solution, we do believe is superior to normal saline, and we would like to question whether normal saline would ever be licensed now, if it was to be developed in the 21st century. The pre-clinical studies would show that it does not maintain electrical or mechanical activity of isolated muscle preparations (2,3). In fact in 1901 in his seminal paper Harvey Cushing called saline ‘poisonous ’(2). The phase 1 studies would show when normal saline is administered to healthy volunteers they subjectively complain of bloatedness, confusion, stomach cramps and essentially feel unwell. Objectively they develop a severe hyperchloraemic acidosis and retain the excess saline and fluid for a longer period than when compared to Hartmann’s solution.(4,5) The Phase 2 studies, and there have been 3 randomised controlled studies comparing the administration of normal saline and balanced solution in high risk patients , again would not support the continued development of normal saline (6,7,8). In fact, 2 of the 3 studies were prematurely halted due to the excess morbidity in what was subsequently discovered to be the normal saline cohort.(6,8) In summary, we reject the concept that it is potentially dangerous to administer Hartmann’s in preference to normal saline in patients with DKA, and we would welcome a study to find the solution to this contentious area (Pun intended). 1. Dhatariya KK. Diabetic Ketoacidosis. BMJ 2007; 334: 1284-5. 2. Cushing H. Concerning the poisonous effect of pure sodium chloride solutions upon the nerve-muscle preparation. Am J Physiol.1901; 6: 77-90 3. Howell W. An analysis of the sodium, potassium, and calcium salts of the blood on the automatic contractions of heart muscle. Am J Physiol.1901; 6: 181-206 4. Williams EL,et al. The effect of intravenous lactated Ringer’s solution versus 0.9% sodium chloride solution on serum osmolality in human volunteers. Anesth Analg 1999;88:999–1003. 5. Reid F, Lobo DN, et al. (Ab)normal saline and physiological Hartmann’s solution: a randomized double-blind crossover study. Clin Sci (Lond) 2003;104:17–24. 6. O’Malley et al. A randomized, double-blind comparison of lactated Ringer's solution and 0.9% NaCl during renal transplantation. Anesth Analg 2005; 100: 1518-24. 7. Waters JH, et al. Normal saline versus lactated Ringer's solution for intraoperative fluid management in patients undergoing abdominal aortic aneurysm repair: an outcome study. Anesth Analg 2001; 93: 817-22. 8. Wilkes NJ, et al. The effects of balanced versus saline-based hetastarch and crystalloid solutions on acid-base and electrolyte status and gastric mucosal perfusion in elderly surgical patients. Anesth Analg 2001;93:811–6. Competing interests: None declared
Maintenance versus Resuscitation Fluids 14 April 2008
Gareth Hynes, F1 Doctor Southampton General Hospital, SO166YD Send response to journal: Re: Maintenance versus Resuscitation Fluids	David Miller, in his thoughts on and contribution to the above discussion, raises a query over why clinicians' choice of infusates is not routinely a closer match to normal plasma values. The answer I think lies with the fact that intravenous fluids are prescribed for (very broadly) two different, though clearly often overlapping, reasons. The first relates to the above situation, where the patient may be intravascularly deplete ('dry') and is at subsequent risk of hypoperfusing their organs. In this situation it seems logical to try to 'fill them up' as quickly as possible, within limits, with an infusate that approaches normal plasma electrolyte levels and isotonicity as closely as possible. Hartmann's, from the above discussion, appears to be the best solution. (The colloid/crystalloid debate aside.) The second relates to patients who are euvolaemic, or 'intravascularly replete', but who cannot take fluids orally, for whatever reason. In this situation the patient will need supplementation for their normal daily losses (± any extra minor losses they may have) - in other words, maintenance fluids. Daily requirements for sodium are 1- 2mmol/kg/day. So a 70kg man will need 70-140mmol/day. Normal Saline provides 154mmol/L, Hartmann's provides 131mmol/L and Dextrose Saline 31mmol/L. If the average patient needs a maintenance of 2-3L of fluid a day, clearly anything other than Dextrose Saline will provide excess sodium, which as noted above is likely associated with adverse outcomes. Normal potassium requirements, on the other hand, are 0.5-1mmol/kg/day, so for the average 70kg man 35-70mmol are needed each day. Hartmann's contains only 5mmol/L, however, so a patient would need 7L per day before reaching the minimal levels of loss they are likely to experience. Adding 2g of KCl (26.6mmol) to each litre of Dextrose Saline, up to 4g per day, would appear more well-matched to a patient's losses. The situation is obviously complicated when the patient is dry (needing filling), doesn't have 'normal plasma values' themselves (rather abnormal plasma values that may need correcting) while also having unknown quantities of electrolyte losses (needing replacement), as well as being nil by mouth - as in Diabetic Ketoacidosis. Hence the protracted discussion above (which I don't pretend to understand in the slightest). Competing interests: None declared