Assessing the degree of extracellular fluid volume contraction in a patient with a severe degree of hyperglycaemia

Supported by an educational grant from A 7-year-old, 20 kg female had a 2 week history of polydipsia, polyuria and a weight loss of ∼2.5 kg. She was drowsy, but easily roused and answered questions appropriately. Blood pressure was 100/60 mmHg, heart rate 148/min, respiratory rate 12/min, capillary refill >3 s and tissue turgor was poor. There was no urine output in the first 2 h (a Foley catheter was inserted). Laboratory data are summarized in Table 1. Based on this information, 98 adult and paediatric medical specialists were asked to assess her degree of extracellular fluid (ECF) volume contraction (Table 2). Laboratory data on admission Albumin was only measured once in plasma; it was 46 g/l 4.5 h after therapy began. Laboratory data on admission Albumin was only measured once in plasma; it was 46 g/l 4.5 h after therapy began. Response to the survey concerning the assessment of the ECF volume in this patient The case synopsis and laboratory data were sent to 150 physicians and trainees. Approximately half of the 98 respondents were from paediatrics and the other half were from adult medicine. In both groups, we consulted endocrinologists, intensivists and nephrologists. Because there was no distinct pattern of response in any subgroup, the results were pooled and are reported for the whole group. Response to the survey concerning the assessment of the ECF volume in this patient The case synopsis and laboratory data were sent to 150 physicians and trainees. Approximately half of the 98 respondents were from paediatrics and the other half were from adult medicine. In both groups, we consulted endocrinologists, intensivists and nephrologists. Because there was no distinct pattern of response in any subgroup, the results were pooled and are reported for the whole group. It is difficult to quantitate the degree of ECF volume contraction on clinical grounds [1–3]. Therefore, laboratory data were examined to help in this regard (Table 1). The combination of a low hydrostatic and high colloid osmotic pressure (high albumin concentration and charge: Donnan effect [4]) implies a greater percentage decline in her interstitial than plasma volume. For simplicity, we shall assume that her ECF volume was reduced from 4 l (20% wt) to 2 l. Her blood volume should be better preserved because of the unchanged RBC volume. Normally, the venous PCO2 is ∼6 mmHg greater than the arterial PCO2. With a low cardiac output, there is a disproportionate rise in venous PCO2 [5]. On admission, her venous PCO2 was 69 mmHg (arterial PCO2: 43 mmHg), implying a very low blood flow rate. No respondent used the venous PCO2 to help assess the circulating volume and its response to therapy. The extreme degree of hyperglycaemia (PGlu) and oliguria together with the modest elevation in plasma creatinine concentration (Table 1) suggested a recent but marked fall in glomerular filtration rate (GFR)— its likely basis was very poor renal perfusion. Perhaps her high blood viscosity decreased flow, especially in the smaller arterioles. The ECF Na deficit was >50% because of ECF volume contraction (>50%) and the PNa of 129 mM (Table 1). Virtually all respondents thought that the Na deficit must be much smaller because of the absence of hypotension (Table 2). Hyperglycaemia is due to either more glucose in the ECF compartment and/or a low ECF volume. Had her GFR been only 20% of normal (0.33 l/h) for the initial 3 h in hospital, she would have excreted almost half of her ECF glucose content [1 l × (110 – 10) mM or 100 mmol vs an ECF glucose content of 220 mmol on admission (2l × 110 mM)]. A 50% decline in ECF volume raised her PGlu 2-fold. Hence, with the same glucose input from her gastrointestinal tract, the very low ECF volume and GFR could virtually quadruple her PGlu from 500 (27.5 mM) to 2000 mg/dl (110 mM). Hyperglycaemia leads to a higher ECF volume because glucose is an effective osmole for skeletal muscle [6]. Using an imaginary redistribution, her ECF was divided into two iso-osmotic solutions because the glucose-containing one will be excreted rapidly when the GFR rises. For this calculation, we assumed that the glucose solution had a PGlu equal to her effective plasma osmolality (Posm) – 368 mOsm/kg H2O in 0.6 l (220 mmol/368 mmol/l). The other 1.4 l contains all the Na with an identical Posm [PNa 184 mM (one-half of 368 mM)]. Hence, this severe degree of hyperglycaemia permitted her to have an ECF volume that was 30% higher than it would have been in the absence of hyperglycaemia [7]. Her low HCO3− content (25 mM × one-half ECF volume) and increased plasma anion gap suggested that she had ‘occult’ metabolic acidosis. The increase in plasma anion gap was magnified by ECF volume contraction. Because her plasma l-lactate− was 0.9 mM, we suspected that she had diabetic ketoacidosis; unfortunately, her plasma β-HB– concentration was not measured. The patient received a bolus of 160 ml saline over 80 min—her venous PCO2 fell to 47 mmHg (Figure 1). Unfortunately, the haematocrit was not measured at this time. Because there was no urine output, there was a 2 ml/min net rise in her ECF volume. This infusion should lower the viscosity of arterial blood to a greater extent, because on first pass it has not reached systemic capillaries. Over the next 120 min, 235 ml of isotonic saline was infused and 200 ml of urine was excreted. Hence, her net fluid gain was only 0.3 ml/min and her venous PCO2 rose to 67 mmHg with no change in arterial PCO2. Of note, her urine output declined appreciably. This venous PCO2 change illustrates the initial benefit of re-expanding the arterial plasma volume and its abrogation by a marked decline in the rate of net fluid addition to this compartment. Time course for changes in haematocrit and venous PCO2. Venous PCO2 is shown as solid squares and the haematocrit is shown as open circles. Note the fall in venous PCO2 at 80 min, a time when intravenous fluid administration was 2 ml/min and there was no urine output. Between 80 and 270 min, the intravenous input was 0.3 ml/min larger than the urine output and the venous PCO2 rose. Two of the important risk factors for cerebral oedema are a rapid rate of infusion of saline and a large fall in the effective Posm [8–10]. Rate of infusion of saline. Because she was awake and responded appropriately to questions, her cerebral blood flow was probably not impaired sufficiently to merit aggressive saline administration (>10 ml/kg/h). Avoiding a fall in the effective Posm. The PGlu will fall due to dilution (infused saline) and glucosuria when the GFR rises [6]. This will lower the effective Posm unless the PNa rises by a similar amount (about one-half the fall in PGlu) [8]. To achieve this constant effective Posm, the tonicity of the intravenous fluids should equal the Posm during oliguria and the urine osmolality (Uosm) during polyuria [11]. Fortunately, the effective Posm and Uosm are similar during a profound glucose-induced osmotic diuresis [12]. Luckily, the osmolality of isotonic saline plus 20–40 mM KCl is similar to this Uosm. Her effective Posm on admission was 368 mOsm/kg H2O and 12 h later it fell to only 359 mOsm/kg H2O (PGlu 27 mM and PNa 166 mM). Later, the decline in PNa should occur gradually to minimize rapid brain cell swelling. ii(i) The magnitude of ECF volume contraction was best revealed by the haematocrit, while the response to therapy was reflected by the venous PCO2 and the change in urine output. i(ii) The rate of saline infusion should be influenced by the urine output and the clinical assessment of the central nervous system status. (iii) Hypotonic fluids should not be given because of the risk of cerebral oedema. Conflict of interest statement. None declared. 1Renal Division, St Michael’s Hospital, University of Toronto, Toronto, 2Division of Endocrinology, Hospital for Sick Children, Toronto, Canada and 3Nephrology Unit and Department of Internal Medicine, Stellenbosch University, Cape Town, South Africa

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.