Share this article on:

Pathophysiologic and Clinical Correlates of Hypophosphatemia and the Relationship with Sepsis and Outcome in Postoperative Patients After Hepatectomy

Giovannini, Ivo; Chiarla, Carlo; Nuzzo, Gennaro

Clinical Aspects

Hypophosphatemia in critically ill and postoperative (p.o.) patients is a multifactorial event, and is also related to severity of illness. This study was conducted to assess pathophysiologic correlates of hypophosphatemia and the simultaneous relationship with clinical events after hepatectomy. A total of 333 measurements were obtained in 59 patients: these were performed preoperatively and at p.o. days 1, 3, and 7 in all patients, and subsequently, until recovery or death, only in those with complications. Measurements included plasma phosphate together with a large number of additional blood chemistries, taking into account primary and associated diseases, events associated with the operation, doses of parenteral substrates, occurrence of sepsis or other p.o. complications, outcome, and a consistent set of complementary variables. Plasma phosphate decreased at p.o. days 1 and 3 (P < 0.001) and returned to a level close to baseline at p.o. day 7. Regression analysis showed that phosphate was related simultaneously to patient age (inversely), levels of creatinine and potassium (directly), and dose of parenteral amino acids (inversely;P < 0.001 for all). Independently of covariation with these variables, there was a decrement in phosphate at p.o. days 1 and 3 that was related specifically to p.o. condition; this decrement had a general component common to all patients, an additional component related to duration of previous hepatic ischemia at surgery, and a further component predictive of the subsequent development of complications (in most cases, sepsis). Plasma phosphate at p.o. day 1 was related inversely to APACHE II score (r2 = 0.4, P < 0.001), and levels lower than 1.5 mg/dL were associated with an almost 4-fold increase in the rate of complications compared with cases with higher phosphate (P < 0.001). The best single variable bridging early evidence of hypophosphatemia to subsequent development of complications was plasma cholesterol, which fell significantly from p.o. day 3 onward in patients with complications compared with those recovering normally (P < 0.01), and in nonsurvivors compared with survivors (P < 0.01). Hypophosphatemia may anticipate clinical evidence of complications by reflecting an early stronger acute-phase response, with shift of phosphate from intra- to extravascular space, or true phosphorus deficiency, which may favor development of complications by impairing high-energy substrate availability for host defense and other cell functions.

Department of Surgery, Hepato-Biliary Unit, and CNR Shock Center, Catholic University School of Medicine, Largo A. Gemelli 8, I-00168 Roma, Italy

Received 23 Jul 2001;

first review completed 13 Aug 2001; accepted in final form 23 Oct 2001

Address reprint requests to Ivo Giovannini, MD, Via Alessandro VII, 45 I-00167 Roma, Italy.

Back to Top | Article Outline


Phosphorus is one of the more abundant substances in the body. However, in inorganic form, including plasma phosphate, it is present in a limited pool in slow equilibrium with the larger organic pool. Inorganic phosphorus is critically important for metabolic processes, in particular those generating energy for life (1), and hypophosphatemia is associated with disruption of physiologic processes and important clinical events. In postoperative (p.o.) and acutely ill patients, hypophosphatemia is common, and has been related to severity of illness and outcome (1–8). However, p.o. changes in plasma phosphate are multifactorial, and it is difficult to identify which of those factors have relevant links with p.o. morbidity and acute illness. After liver resection, most factors may become simultaneously present, and thus the operation provides an ideal model to assess biochemical, pathophysiologic, and clinical correlates of hypophosphatemia. This study has been carried out to perform such an assessment over a large group of patients undergoing hepatectomy.

Back to Top | Article Outline


Liver resections were performed in 26 female and 33 male patients. Age was 57 ± 13 yrs (mean ± SD), body weight was 70 ± 13 kg, body surface area was 1.77 ± 0.17 m2, and body mass index (weight/height2) was 25.3 ± 4.1 kg/m2. Indications were primary malignancy in 26 patients (hepatocarcinoma in 16, cholangiocarcinoma in seven, carcinoma of gallbladder in two, and hepatoblastoma in one), hepatic secondaries in 17 (from colorectal cancer in 10, and from other sources in seven), and benign lesions in 16 (focal nodular hyperplasia in five, hemangioma in two, adenoma in two, and congenital bile duct dilation, benign or hydatid cysts, or other lesions in the remainder). Fifteen patients had liver cirrhosis (10 from type C hepatitis, two from type B, and three from type C plus B); all were in Child A state and good general condition. According to ASA classification, 35 patients were in class I, four were in class II, 19 were in class III, and one was in class IV.

Hepatectomies were all performed by the same senior surgeon (G.N.) and consisted of 25 minor resections (12 resections of one liver segment and 13 of two segments) and 34 major resections: the latter included four resections of three segments, one of four segments, 13 left hepatectomies (one extended to segment I, and two to segment V and VIII), and 16 right hepatectomies (five extended to segment I and/or to IV). There were nine associated bowel operations (resections for primary malignancies or Roux-en-Y biliary reconstructions).

Forty-three patients recovered normally. Eleven had nonlethal complications: two had intraabdominal sepsis, one had pulmonary sepsis, four had a biliary fistula (transiently associated with sepsis in 3), three had liver insufficiency (defined by two or more of the following: bilirubin > 12.0 g/dL in the absence of biliary obstruction, prothrombin activity < 35%, relevant ascites, encephalopathy), and one had ventilatory insufficiency with prolonged mechanical ventilation. Diagnosis of sepsis was based on previously described criteria (9) involving the presence of a systemic septic response, identification of the source of infection, and positive cultures from blood, pus (drained from abdominal sources), bile (in cases with biliary fistula and/or cholangitis), and sputum (in the case of pulmonary infection). Nonsurvivors included two patients observed during the prospective period of the study and also three observed outside this period. This inclusion only improved the significance of results in the small number of nonsurvivors, without bias, as the pattern of death was similar in all cases (development of systemic sepsis, combined with liver insufficiency or acute respiratory distress syndrome, progressing in all cases to irreversible multiple organ dysfunction syndrome). All five patients had primary or secondary malignancy; two had also liver cirrhosis. Patients did not receive aluminum-containing antacids.

This patient population provided a continuous distribution of observations, from minor to superextended surgical procedures (and from minor to extremely severe p.o. illness, or preterminal state), which was perfectly suited to assessing correlates of hypophosphatemia over a wide range of pathophysiologic abnormalities.

According to our protocol for hepatectomies, in the early p.o. period, all patients received low-dose parenteral glucose (1.9 ± 1.4 g/kg/day), with amino acids (0.9 ± 0.6 g/kg/day, preferentially 35%–50% branched-chain) to support liver regeneration, up to p.o. day 3. Thereafter, parenteral regimen was continued only if oral feeding could be not resumed, and reached full-dose parenteral nutrition, including fat (50% medium-chain and 50% long-chain triglycerides), in cases with complications and prolonged illness. There was no exogenous phosphate supplementation up to p.o. day 3. Subsequently, administration of parenteral phosphate (fructose 1-6 diphosphate, Esafosfina®, Biomedica Foscama, Ferentino, Italy, or potassium phosphate) was started if normal feeding could not be resumed, and was maintained as a common component of parenteral nutrition (which contained phosphate also in fat emulsions). It was withheld only in cases with hyperphosphatemia from renal insufficiency. Doses of phosphate ranged from 20 to 50 mM/day.

Altogether, 333 venous blood measurements were obtained. These were performed according to a clinical routine, preoperatively and then postoperatively at days 1, 3, and 7 in all patients; in those with complications, measurements were continued until recovery or death. The following variables were considered: demographic data, plasma phosphate concentration, sodium, potassium, calcium, chloride, magnesium, blood urea nitrogen, creatinine, urate, glucose, cholesterol, triglycerides, albumin, total protein, albumin/globulin ratio, fibrinogen, alanine aminotransferase (ALT), aspartate aminotransferase (AST), peak ALT and AST, alkaline phosphatase, gamma-glutamyl-transpeptidase, amylase, total and indirect bilirubin, peak bilirubin, prothrombin activity, hematocrit, hemoglobin, blood cell count, body temperature, number of resected liver segments, occurrence of an associated bowel operation, duration of the operation and of eventual hepatic pedicle clamping (normothermic liver ischemia, mean duration 47 ± 29 min, range 5–110, used in 39 patients), blood loss, administration of crystalloids, colloids and blood transfusions at surgery, and occurrence of sepsis or other p.o. complications, with or without death. Occurrence of neoplastic disease, previous chemotherapy, cirrhosis, and diabetes were also considered as independent variables. Magnesium was not measured constantly in all patients (n = 151). APACHE II score (10) at p.o. day 1 was recorded in 53 patients. Statistical analysis involved the use of Student's t test, the chi-square test, and regression analysis. The latter was based on least-square regressions, with analysis of residuals, skewness, and kurtosis control, and “simplest best fit” procedures selecting the simplest possible regressions controlling the largest possible variability of plasma phosphate, based on Mallows' Cp criteria.

This study was approved by our Institutional Human Use Committee and informed consent was obtained from either the patient or next of kin.

Back to Top | Article Outline


Compared with preoperative phosphate (3.1 ± 0.6 mg/dL, mean ± SD), p.o. levels fell significantly at p.o. day 1 (2.1 ± 0.8, P < 0.01), fell further at p.o. day 3 (1.7 ± 0.8, P < 0.01), and returned to a level not significantly different from baseline at p.o day 7 (2.8 ± 1.1).

Regression analysis showed that preoperatively there was a weak inverse relationship between phosphate and age, weak direct relationships between phosphate and sodium and creatinine and potassium (P < 0.01 for all) and a trend for lower phosphate in patients with liver cirrhosis (P = 0.05), and/or decreased albumin/globulin ratio (P < 0.01). There were no significant relationships with other variables, including subsequent occurrence of complications or outcome.

Analysis of p.o. measurements reconfirmed an inverse relationship between phosphate and age, and showed direct relationships with creatinine, blood urea nitrogen, uric acid, potassium, magnesium, albumin/globulin ratio, ratio of cholesterol to preoperative value, amylase, lymphocytes, platelet count, and triglyceride level (r2 = 0.25 to 0.06, P < 0.001 for all). There were also inverse relationships with parenteral doses of amino acids and glucose (P < 0.001 for both); the stronger correlation was with doses of amino acids (r2 = 0.14), whereas doses of glucose lost significance when both variables were considered together in the same regression. Correlation with doses of amino acids was supported more by the branched-chain than by the non-branched-chain component (P < 0.01). All these correlations had similar coefficients when considering measurements performed in the early p.o. period (p.o. days 1–3), those performed at a later time, or all pooled together. There were no other significant relationships with other variables, although within a reduced number of complementary measurements excluded from the main results, there was a relatively strong inverse relationship between p.o. phosphate and heart rate (r2 = 0.20, P < 0.001).

Regression analysis reconfirmed at p.o. day 1 a significant decrement in phosphate (P < 0.001), which was independent of covariation of phosphate with the other variables, except that the fall was significantly larger (P < 0.001) in patients who subsequently developed complications or died. At p.o. day 3, a similar decrement was reconfirmed, which was similarly larger in patients who subsequently developed complications (at this time, some of them were already showing initial signs of sepsis). In addition, the decrement at p.o. day 3 was significantly larger in patients who had undergone liver ischemia at previous surgery, and was related to duration of ischemia (P < 0.001). Regression analysis showed further that after p.o. day 3, phosphate was slightly but not significantly lower than baseline (always by accounting simultaneously for covariation with other variables). By this time, however, plasma phosphate could have been influenced by exogenous phosphate supply, which was started after p.o. day 3 in patients with hypophosphatemia who were unable to resume normal feeding, and was maintained thereafter as a normal component of parenteral nutrition in patients requiring more prolonged support. Our study was not designed to assess the impact of phosphate supply, as it did not include patients in equivalent conditions receiving or not receiving phosphate; rather, phosphate was administered only later in the clinical course in patients still requiring parenteral nutrition to maintain phosphate balance in the presence of prolonged deprivation. There were no differences in plasma phosphate in cases receiving fat with parenteral nutrition after p.o. day 3 compared with those who were not.

The analysis showed finally that the best simultaneous correlates of plasma phosphate, in all measurements pooled together, were age, creatinine, potassium, dose of amino acids, and the decrements at p.o. days 1 and 3, including the components specifically related to duration of liver ischemia and development of complications. These variables explained together 69% of the variability of phosphate in a multiple regression (Table 1), which was the best descriptor of the evolution of phosphate along the clinical course. Inclusion of any other variables or events in the regression (even inclusion of death), did not significantly improve the r2. In fact, inclusion of albumin/globulin ratio and lymphocyte count brought the total r2 closer to 0.80, but the regression lost strength, as the number of variables was not supported by a sufficiently high n and r2. Also, substitution of potassium with magnesium yielded a higher r2, but with a much lower n, as magnesium was not measured consistently. Finally, inclusion of triglycerides and amylase apparently increased the r2, but this was not supported by skewness criteria (being mostly due to large simultaneous increases in plasma triglycerides, amylase, and phosphate in cases with end-stage sepsis, multiple organ dysfunction syndrome, and renal failure, and not to a real correlation).

Table 1

Table 1

Grouping of patients according to plasma level of phosphate at p.o. day 1 showed that about three-quarters of cases with phosphate lower than 1.5 mg/dL subsequently developed complications (Table 2), including three nonsurviving patients. This was unrelated to differences in potentially relevant factors (demographics, body size indices, magnitude and duration of operation, and need for transfusion and fluid balance at surgery) apart from tendency to occur more often in patients with cirrhosis (n.s.) or malignancy (P < 0.05). Cirrhosis and malignancy were associated with a trend towards a higher rate of complications, independent of phosphate levels; however, the trend was not significant in cirrhosis and reached borderline significance (P < 0.05) in malignancy. Early p.o. hypophosphatemia was unrelated to preoperative and early p.o. values of prothrombin activity, plasma bilirubin, albumin, ALT, and AST in all cases and in cirrhotic patients. Malnutrition was not an important factor because no patient with severe malnutrition underwent surgery, and there was no relationship between body size indices and rate of complications. Ratio of actual-to-ideal body size was less than 0.90 in five cases only (all of whom were judged to be in good condition in spite of the body weight criteria): among these, there was major p.o. hypophosphatemia in only one case, who subsequently developed a complication. Also, hypophosphatemia at p.o. day 1 was unrelated to the presence of cardiovascular or chronic respiratory diseases, body size indices, and ASA score, whereas it was inversely related to the APACHE II score (r2 = 0.40, P < 0.01) in the 53 patients in whom the score could be calculated, including four of the five nonsurvivors and 10 of the 11 survivors with complications.

Table 2

Table 2

Grouping of patients according to plasma level of phosphate at p.o. day 3 provided results similar to those found at p.o. day 1, but these were less specific, as phosphate level was influenced also by length of hepatic ischemia.

Analysis of the data provided additional interesting details. The relationship between duration of liver ischemia and decrease in phosphate was a unique feature of measurements performed at p.o. day 3. This decrease was simultaneously related to decrease in plasma cholesterol (P < 0.01), and was unrelated or poorly related to all other variables, including the use of continuous vs. intermittent liver ischemia. Analysis of other changes associated with duration of ischemia showed only weak direct relationships with ALT and AST at p.o. day 1 (not at p.o. day 3 or later) and with peak bilirubin.

Analysis of the correlations with nadir phosphate value (the lowest value reached in each patient between p.o. days 1 and 3) did not yield additional relevant information, except that nadir value was weakly and inversely correlated with duration of the operation (P < 0.01).

Autocorrelations between phosphate levels on different days showed weak direct relationships between preoperative values and values at p.o. days 1 and 3 (P < 0.01 for both). Correlations between the fall in phosphate at p.o. days 1 and 3 and changes in other variables at p.o. day >3, showed significant relationships with late modifications of albumin/globulin ratio, cholesterol, and blood cell count (P < 0.001), which were characteristic of sepsis, thus reconfirming the predictive value of the early fall in phosphate with regard to occurrence of subsequent complications. Among these modifications, particularly relevant were those of plasma cholesterol, the best single variable characterizing evolution following early p.o. hypophosphatemia. At p.o day 1, cholesterol was unrelated to APACHE II score, and was not significantly different in patients with or without complications. At p.o. day 3, it became significantly lower in patients with complications and/or earlier more severe hypophosphatemia compared with the remainder (P < 0.01), and in nonsurvivors compared with survivors (P < 0.01). Modifications after p.o. day >3 included maintenance of more severe hypocholesterolemia in patients with complications who subsequently died compared with those who did not (63 ± 13 vs. 95 ± 41;P < 0.001), with a pattern of death characterized invariably by extremely severe and persistent hypocholesterolemia.

Finally, phosphate at p.o. day 1 correlated inversely with length of p.o. hospital stay in survivors (P < 0.001), as an obvious consequence of the greater rate of complications associated with early hypophosphatemia.

Back to Top | Article Outline


Our study provides an assessment of the multifactorial nature of hypophosphatemia after hepatectomy, and of its clinical and pathophysiologic correlates. We found that phosphate level, besides being related inversely to age, varied along the clinical course on average by 0.86 mg/dL per unit change in creatinine, by 0.39 mg/dL per unit change in potassium, and decreased by 0.34 mg/dL per unit increase in dose of parenteral amino acids (Table 1). When assessing clinical relevance of hypophosphatemia, it is important to account for covariation of phosphate with these variables because changes in phosphate may reflect p.o. modifications of renal function, ion balances, and substrate doses without any link with other major clinical events (1,11). By accounting for such changes, our study has also assessed a fall in phosphate at p.o. days 1 and 3 that was specifically related to postoperative condition. There were three components of this fall: a general component common to all patients (−0.5 mg/dL), an additional component, evident only at p.o. day 3 and related to duration of previous hepatic ischemia (−0.012 mg/dL per minute ischemia), and a third component that was predictive of subsequent development of complications, with or without death (−0.7 mg/dL).

These results, obtained prospectively, substantiate the existence of a relationship between early p.o. hypophosphatemia and increased rate of complications, as suggested previously by smaller, retrospective studies (12,13). Different from those studies, our findings emerge from a very extensive quantitative assessment of simultaneous correlates of plasma phosphate. Furthermore, the magnitude of the operation varied largely among different patients (and was not found to be a relevant factor). We also outlined, in nonsurviving patients, the pattern of death on the basis of more severe and persistent hypocholesterolemia developing after early p.o. hypophosphatemia.

The mechanism by which severe hypophosphatemia anticipates clinical evidence of a complication is unclear. However, the strong relationship found with APACHE II score suggests that severe hypophosphatemia is the humoral marker of a more intense acute-phase response, preceding overt manifestation of the complication. Indeed, plasma phosphate may be lowered in such conditions by respiratory alkalosis (a common feature of acute-phase response) and the effect of hormones (insulin, glucagon, cortisol, and catecholamines), interleukin 6, and other proinflammatory cytokines (1,6,8,14–17). Early increase in level of interleukin 6 after hepatectomy has been specifically associated with subsequent development of complications (18). A very strong acute-phase response is elicited by sepsis, and sepsis prevailed largely among lethal and nonlethal complications in our patients. A septic component may also be present in p.o. liver insufficiency, when hepatocyte dysfunction is paralleled by Kupffer cell dysreactivity, with inadequate clearance of bacteria and bacterial products from portal blood and enhancement of acute-phase and other inflammatory responses (19–23). In biliary fistula without sepsis, exposure of tissues and peritoneum to bile, before external drainage of the fistula, could be the triggering event for a stronger acute-phase response. In these instances, hypophosphatemia should depend mostly on a shift of phosphate from intra- to extravascular space. Conversely, renal wasting of phosphate mediated by hypocalcemia, through stimulation of parathormone release, is not supported by the lack of correlation between phosphate and calcium; at any rate, hypophosphatemia does not seem to be associated with increased parathormone levels in post-traumatic states, or with increased renal wasting of phosphate after burns (2,6,8,24).

In addition to a shift from intra- to extravascular space, true phosphorus deficiency may also be involved after hepatectomy. Plasma phosphate belongs to the small body pool of inorganic phosphorus, and development of hypophosphatemia is normally buffered by the larger organic pool (1). The rate of buffering may be exceeded by acutely increased metabolic demands, especially after severe phosphate depletion. This may occur at the second to third p.o. day after hepatectomy, when demand for liver regeneration and consumption of ATP by rapidly dividing liver cells become maximized (12,13,25,26). Correlation between hypophosphatemia and hypocholesterolemia at p.o. day 3 in our patients is consistent with maximum substrate demand for regenerative and reparative processes (27,28). Maximization of demand may unmask a state of phosphorus depletion, such as that related to previous liver ischemia, thus explaining the correlation found also at p.o. day 3 between degree of hypophosphatemia and duration of ischemia. Besides hepatectomy, other operations involving extensive tissue ischemia are associated with a more severe fall in phosphate shortly after p.o. day 1 (5). Postoperative hypermetabolism may further amplify the need for phosphorylated intermediary compounds of energy metabolism, and has been related to development of hypophosphatemia in liver transplantation and burn injury (8,29,30). Finally, parenteral substrates also contribute to raising the demand for phosphorus (1), and our study has shown an effect of amino acids on hypophosphatemia greater than that of glucose, which is consistent with the greater stimulation of metabolic processes induced by amino acids in acute conditions (9,31,32).

If phosphorus deficiency limits generation of intracellular ATP and other key processes, serious consequences may ensue, including death in extreme cases (such as in refeeding after prolonged starvation) (1). It is also known to impair chemotactic, phagocytic, and bactericidal activity of granulocytes by limiting availability of ATP and energy, of other phosphate compounds necessary for cell function, and by limiting the capability to maintain low intracellular calcium (1). The role of phosphorus deficiency as a cause of clinically relevant immune dysfunction is not yet totally defined, however, it may contribute to explaining the high incidence of sepsis in our hypophosphatemic patients. Respiratory muscle dysfunction is known to depend on shortage of free phosphate, and could have mediated the development of ventilatory insufficiency in one of our patients (1,33), whereas the relationship between the shortage of free phosphate and subsequent development of liver failure remains unclear (34).

In conclusion, our study cannot assess exactly whether the pathophysiologic mechanism linking severe p.o. hypophosphatemia to subsequent development of sepsis, of other complications, or death after hepatectomy involves a stronger stress response, phosphorus deficiency, or both factors together. However, our results expand the information available from previous studies (1,12,13) and have clinical implications that may be relevant for patient monitoring and treatment: 1) with the modern tendency to simplify p.o. management after hepatectomy, especially after minor resections, occurrence of severe p.o. hypophosphatemia offers an helpful landmark to anticipate challenging events and the need for greater support and care; 2) the best humoral variable bridging occurrence of early p.o. hypophosphatemia to subsequent development of complications is severe hypocholesterolemia, which becomes more extreme and persistent in patients evolving to multiple organ dysfunction syndrome and death; 3) our data might support the need for routine exogenous supply of phosphorus, starting perhaps during surgery, or immediately after (always avoiding hyperphosphatemia from unrestrained doses) (1); however, it is not clear whether such supply would really avert development of complications, as suggested by a previous retrospective study (12) not validated prospectively. Additional investigations are needed not only to clarify this issue, but also to explain individual exceptions to the link between early p.o. hypophosphatemia and subsequent development of complications (Table 2) and to better assess the relationship with death.

Back to Top | Article Outline


The authors acknowledge the helpful contribution of Mr. Maurizio Cianfanelli to the development of this work, and the editing assistance of Miss Helen Raiswell.

Back to Top | Article Outline


1. Knochel JP: Phosphorus. In Shils ME, Olson JA, Shike M, Ross AC (eds): Modern Nutrition in Health and Disease. Baltimore, MD: Williams & Wilkins, 1999, pp 157–167.
2. Finsterer U, Betz J, Braun S, Beyer A, Jensen U, Kellermann W: Metabolism of phosphate and calcium after severe accidental trauma. Scand J Clin Lab Invest 165(Suppl):117–122, 1983.
3. Lovén L, Larsson J, Lennquist S, Liljedahl SO: Hypophosphatemia and muscle phosphate metabolism in severely injured patients. Acta Chir Scand 149:743–749, 1983.
4. Lovén L, Nordstrom H, Lennquist S: Changes in calcium and phosphorus and their regulating hormones in patients with severe burn injuries. Scand J Plast Reconstr Surg 18:49–53, 1984.
5. Andersen PT, Nielsen LK, Moller-Petersen J, Henneberg EW, Egeblad K: Severe hypophosphatemia following elective abdominal aortic bypass grafting. Acta Chir Scand 153:641–646, 1987.
6. Daily WH, Tonnesen AS, Allen SJ: Hypophosphatemia: incidence, etiology and prevention in the trauma patient. Crit Care Med 18:1210–1214, 1990.
7. Berger MM, Rothen C, Cavadini C, Chiolero RL: Exudative mineral losses after serious burns: a clue to the alterations of magnesium and phosphate metabolism. Am J Clin Nutr 65:1473–1481, 1997.
8. Dickerson RN, Gervasio GM, Sherman JJ, Kudsk KA, Hickerson WL, Brown RO: A comparison of renal phosphorus regulation in thermally injured and multiple trauma patients receiving specialized nutrition support. J Parenter Enteral Nutr 25:152–159, 2001.
9. Giovannini I, Chiarla C, Boldrini G: The relationship between oxygen extraction and venous pH in sepsis. Shock 8:373–377, 1996.
10. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 13:818–829, 1985.
11. Crook MA: Hypophosphatemia and hypokalemia in patients with hypomagnesaemia. Br J Biomed Sci 51:24–27, 1994.
12. George R, Shiu MH. Hypophosphatemia after major hepatic resection. Surgery 111:281–286, 1992.
13. Buell JF, Berger AC, Plotkin JS, Kuo PC, Johnson LB: The clinical implications of hypophosphatemia following major hepatic resection and cryosurgery. Arch Surg 133:757–761, 1998.
14. Mostellar ME, Tuttle EP: Effect of alkalosis on plasma concentration and urinary excretion of inorganic phosphate in man. J Clin Invest 43:138–149, 1964.
15. Massara F, Camanni F: Propanolol block of adrenalin-induced hypophosphataemia in man. Clin Sci 38:245–250, 1970.
16. Giovannini I, Chiarla C, Boldrini G, Castagneto M: Impact of fat and glucose administration on metabolic and respiratory interactions in sepsis. J Parenter Enteral Nutr 13:141–146, 1989.
17. Barak V, Schwartz A, Kalickman I, Nisman B, Gurman G, Shoenfeld Y. Prevalence of hypophosphatemia in sepsis and infection: the role of cytokines. Am J Med 104:40–47, 1998.
18. Chandra Das B, Isaji S, Kawarada Y: Analysis of 100 consecutive hepatectomies: risk factors in patients with liver cirrhosis or obstructive jaundice. World J Surg 25:266–273, 2001.
19. Pardy BJ, Dudley HAF: Post-traumatic pulmonary insufficiency. Surgery 144:259–269, 1977.
20. Siegel JH, Giovannini I, Coleman B, Cerra FB, Nespoli A: Pathologic synergy in cardiovascular and respiratory compensation with cirrhosis and sepsis. Arch Surg 117:225–238, 1982.
21. Katz S, Grosfeld JL, Gross K, Plager DA, Ross D, Rosenthal RS, Hull M, Weber TR: Impaired bacterial clearance and trapping in obstructive jaundice. Ann Surg 199:14–20, 1984.
22. Liu P, McGuire GM, Fisher MA, Farhhod A, Smith CW, Jaeschke H: Activation of Kupffer cells and neutrophils for reactive oxygen formation is responsible for endotoxin-enhanced liver injury after hepatic ischemia. Shock 3:56–62, 1995.
23. Fox ES, Tracy Jr: TF Alterations in tumor necrosis factor-α expression by hepatic macrophages following acute cholestatic liver injury. Shock 5:112–115, 1996.
24. Drop LJ: Ionized calcium, the heart and hemodynamic function. Anesth Analg 64:432–451, 1985.
25. Fausto N: Hepatic regeneration. In Zakim D, Boyer TD (eds): Hepatology. Philadelphia, PA: Saunders, 1990, pp 49–64.
26. Murphy EJ, Brindle KM, Rorison CJ, Dixon RM, Rajagopalan B, Radda GK: Changes in phosphatidylethanolamine metabolism in regenerating rat liver as measured by 31P-NMR. Biochim Biophys Acta 1135:27–34, 1992.
27. Giovannini I, Boldrini G, Chiarla C, Giuliante F, Vellone M, Nuzzo G: Pathophysiologic correlates of hypocholesterolemia in critically ill surgical patients.
28. Nuzzo G, Giovannini I, Giuliante F, Chiarla C, Boldrini G, Marrocco M: La rigenerazione epatica nel paziente a rischio, non cirrotico. Proceedings of the Seventh Congress of the Italian Society for Surgical Pathophysiology, Padua, Italy, 1994.
29. Chiarla C, Giovannini I, Castagneto M: Pathophysiologic patterns after liver transplantation. Proceedings of the 24th Congress of the Italian Society of Organ Transplantation, Roma, Italy, 1997.
30. Yu YM, Tompkins RG, Ryan CM, Young VR: The metabolic basis of the increase in energy expenditure in severely burned patients. J Parenter Enteral Nutr 23:160–168, 1999.
31. Chiarla C, Siegel JH, Kidd S, Coleman B, Mora R, Tacchino R, Placko R, Gum M, Wiles III, CE Belzberg H, Rivkind A: Inhibition of post-traumatic septic proteolysis and ureagenesis and stimulation of hepatic acute-phase protein production by branched-chain amino acid TPN. J Trauma 28:1145–1172, 1988.
32. Giovannini I, Chiarla C, Boldrini G, Castiglioni GC, Castagneto M: Calorimetric response to amino acid infusion in sepsis and critical illness. Crit Care Med 16:667–670, 1988.
33. Aubier M, Murciano D, Lecocguic Y, Viires N, Jacquens Y, Squara P, Pariente R: Effect of hypophosphatemia on diaphragmatic contractility in patients with acute respiratory failure. N Engl J Med 313: 420–424, 1985.
34. Knochel JP: Does hypophosphatemia play a role in acute liver failure? Hepatology 9:504–505, 1989.

Plasma phosphate; liver resection; acute-phase response; critical illness; surgical infections; hypocholesterolemia; morbidity; mortality

© 2002 Lippincott Williams & Wilkins, Inc.