Hepatic Resection-Related Hypophosphatemia Is of Renal Origin as Manifested by Isolated Hyperphosphaturia

To the Editor:

Seven years ago, we prospectively investigated early postsurgical modifications of the calcium-PTH axis.¹ Following the article by Salem and Tray,² we retrospectively reexamined patients who underwent hepatic resection in our study and for whom we had data on serum phosphate, ionized calcium, HCO3−, pH, and intact PTH determined before, during the first and second hours of surgery, and serially thereafter on postoperative days 1, 2, 3, 5, and eventually 7. Fractional phosphate excretion could be quantified before surgery from 12-hour urinary estimation and from 24-hour urinary estimation on postoperative days 1, 2, 3, 5, and eventually 7. The results are expressed as mean ± SEM.

Nine patients (7 women and 2 men), with a mean age of 42.5 ± 3.1 years, underwent liver resection for hepatic metastases secondary to colon cancer (n = 4), focal nodular hyperplasia (n = 3), giant hemangioma (n = 1), or hydatid cyst (n = 1). Hepatectomy mean duration was 205 ± 15 minutes, and mean intraoperative bleeding was 750 ± 171 mL. Six of 9 patients needed pedicle clamping (mean duration, 35.3 ± 5.7 minutes). All patients were extubated at the end of surgery. Five of 9 patients were admitted to the intensive care unit for 24 hours. Oral diet was allowed on postoperative day 3 for the majority of patients. No major complications occurred postoperatively, but 3 patients presented with transient fever associated with atelectasis. A decrease in serum phosphate was observed as early as the first and second hours during liver resection, from 1.25 ± 0.07 mmol/L (normal, 0.70–1.30 mmol/L) to 0.99 ± 0.04 and 1.02 ± 0.04, respectively. Serum phosphate reached a nadir on postoperative day 2 (0.49 mmol/L ± 0.10, P < 0.001), which represented a decrease of 60.8%. Serum phosphate at its nadir was inversely correlated with the number of liver segments removed (P < 0.01), but not with pedicle clamping duration (P > 0.5). Fractional phosphate excretion increased significantly on postoperative days 1 and 2, from 12.4% ± 1.2% (normal, 10%–15%) to 26.8% ± 1.1% and 35.6% ± 5.6%, respectively (P < 0.05), which represents a 116% to 187% increment. Fractional excretion of phosphate peaked on postoperative day 2 where it was correlated with the number of segments removed (P < 0.01) and inversely correlated with serum phosphate (r = −0.49; P < 0.001), but not with the duration of pedicle clamping (P = 0.06). Rapid changes in serum Ca²⁺ were noted at the beginning of the surgical procedure with a significant fall (P < 0.01) from 1.22 ± 0.009 mmol/L (normal, 1.16–1.29 for pH: 7.4) to 1.03 ± 0.01 mmol/L and 0.98 ± 0.01 mmol/L on the first and second hours (nadir) of surgery, respectively, before returning to its baseline level 36 hours after surgery. Intact PTH rose significantly (P < 0.01) during hepatectomy, from 3.4 ± 0.4 pmol/L (normal, 1.4–6.8 pmol/L) to 13.5 ± 2 and 25.5 ± 4.2 (7-fold increase) on the first and second hours of surgery, respectively, and declined thereafter to 9.7 ± 1.9 pmol/L 24 hours after surgery and to 4.7 ± 0.7 pmol/L (normal) at 36 hours. Intact PTH was not correlated with serum phosphate (r = −0.17; P > 0.2) but was inversely correlated with Ca²⁺ (r = −0.61; P < 0.0001) and significantly associated with fractional phosphate excretion, albeit with a limited degree of correlation (r = 0.34; P < 0.05). During the first 2 postoperative days, fractional phosphate excretion and PTH varied inversely, and the fractional phosphate excretion peaked when intact PTH returned to its normal values. Only 1 patient needed substitution for frank hypophosphatemia (0.25 mmol/L).

Hypophosphatemia after major hepatectomy was initially reported by Keushkerian and Wade in 1984,³ but very few data have been published thereafter. Several physiologic mechanisms were evoked; however, dramatic urinary phosphate wastage was demonstrated recently by Salem and Tray.² Chronologically, lowered phosphatemia correlated better with urinary phosphate leakage than with an hypothetic process, such as phosphate shifting into cells like hepatocytes. Liver regeneration lasts no less than 4 weeks⁴ and therefore cannot be a credible reason for transient hypophosphatemia that usually persists for 2 to 3 days. Since the mid-1990s, hypophosphatemia has been observed early after liver resection in our department. Phosphate monitoring is routinely performed with phosphate substitution when the serum level drops to less than 0.5 mmol/L. Like Salem and Tray,² we have observed similar postoperative variations of serum phosphate and its fractional excretion, which were progressively resolved by postoperative days 5 and 7. Salem and Tray stated that “posthepatic hypophosphatemia and its concomitant renal phosphate leak may well reflect the action of an as yet unidentified circulating agent responsible for phosphate homeostasis.”² Similar phosphate metabolism disorders with hypophosphatemia and deficit of phosphate renal handling were noted in other pathologic settings, mainly X-linked hypophosphatemic rickets, autosomal-dominant hypophosphatemic rickets, and oncogenic osteomalacia.⁵ The study of these disorders has resulted in the discovery of a number of proteins, which were named phosphatonins, fibroblast growth factor-23 (FGF-23), secreted frizzled related protein-4 (sFRP-4), and matrix extracellular phosphoglycoprotein (MEPE).⁵ However, by controlling renal handling of phosphate in proximal tubules, PTH plays a central role in phosphate homeostasis.⁵ Salem and Tray did not find any significant variation of PTH and FGF-23 using a carboxyl-terminal FGF-23 assay in 2 and 4 patients, respectively. The prospective and systematic monitoring we performed in our study showed PTH level increased 7-fold and peaked in the second surgical hour before returning to its baseline level by 36 hours. Intact PTH variation appeared to be inversely correlated with serum Ca²⁺ as an acute response to its changes during surgery. The plasma concentrations of Ca²⁺ and inorganic phosphate are primarily under the control of PTH and vitamin D.⁶ Variations in Ca²⁺ are controlled by a specific Ca²⁺-sensing protein in the cell membrane of parathyroid cells, leading to the desired changes in PTH secretion.⁷ Through its effects on bone and intestinal absorption, PTH tends to augment phosphate entry into the extracellular fluid. As well, PTH is inclined to reduce proximal tubular phosphate reabsorption, resulting in increased fractional excretion. However, the urinary effect usually predominates in patients with relatively normal renal function, as PTH tends to lower the phosphatemia.⁵ Knowing the rapid effect of PTH on proximal tubular phosphate reabsorption,⁵ hyperphosphaturia after liver resection appears to be mediated by factors other than PTH. PTH and fractional excretion of phosphate were chronologically uncorrelated; their respective variations were reversed during the first 48 postoperative hours.

Hypophosphatemia with hyperphosphaturia appears to be a constant postoperative disorder after hepatectomy. Rapid hypocalcemia during surgical procedures is associated with increased PTH as an acute response. However, PTH is chronologically unrelated with later urinary phosphate leakage. Therefore, posthepatectomy hypophosphatemia, as a consequence of elevated fractional excretion of phosphate, appears to be mediated by phosphaturic factors other than PTH. The role of phosphatonins (FGF-23, MEPE, and sFRP-4) in this setting was not yet well studied, an aspect that we are currently investigating.

Otmane Nafidi, MD*
Raymond Lepage, PhD^‡
Real W. Lapointe, MD*
Pierre D’Amour, MD^†
Department of *Surgery, ^†Biochemistry,
and ^‡Medicine
Centre Hospitalier de l’Université de
Montréal Hôpital St. Luc
Montréal, Québec, Canada
raymond.lepage.chum@ssss.gouv.qc.ca