Delayed Hyponatremia Following Surgery for Pituitary Adenomas: An Under-recognized Complication
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.280637 PMID: 32189698
Source of Support: None, Conflict of Interest: None
Keywords: Cerebral salt wasting, delayed hyponatremia, diabetes insipidus, pituitary adenoma, surgery, syndrome of inappropriate antidiuretic hormoneKey Messages: Delayed hyponatremia is usually asymptomatic and occurs around the 7th post-operative day following pituitary surgery.
Delayed hyponatremia is an under-recognized complication seen following pituitary surgery. It is defined as serum sodium <135 mmol/L after the third post-operative day., Delayed hyponatremia occurs in 9%–23% of patients after pituitary surgery [Table 1]. These patients are usually discharged after 72 h and this complication occurs after they leave the hospital. The consequences can be devastating if they are not forewarned about the symptoms of hyponatremia and the need to seek immediate medical attention.
The first publication about this entity came out in 1995, but most neurosurgeons and physicians are either unaware of this complication or do not accord it the importance it deserves. The main aim of this study was to draw attention to this complication by documenting its incidence, presentation, management, and outcome. We looked for possible risk factors which could predict the occurrence of delayed hyponatremia in patients undergoing pituitary surgery.
This was a retrospective study of all patients with nonfunctioning pituitary macroadenomas who underwent surgery from 2007 to 2016. Patients with pre-existing diabetes insipidus (DI) were excluded. This study comprised of 222 patients, and included 139 males and 83 females. The median age of these patients was 45 years (range 17–74 years). All underwent pituitary surgery via the trans-sphenoidal route; 2 also had trans-cranial surgery at the same sitting.
All patients were evaluated and treated using a standard protocol which included intravenous hydrocortisone 25 mg 6 hourly for the first 24 h. Intravenous hydrocortisone was slowly tapered and withdrawn after 72 h and these patients were started on oral prednisolone 10 mg once daily. Over a period of two months prednisolone was gradually tapered to 5 mg on alternate day. They came back for their first follow up visit after three months and under close supervision, prednislone was withdrawn for 48 hours and their 8AM cortisol was retested. Only those with low cortisol levels (<3.5 μg/dL) were asked to continue oral prednisolone long term. Those with borderline cortisol levels (3.5–10 μg/dL) were advised steroid cover during stress situations.
Delayed hyponatremia was defined as serum sodium <135 mmol/L occurring after the third post-operative day. Hyponatremia was classified as mild (134–130 mmol/L), moderate (129–125 mmol/L), and severe (<125 mmol/L).
Patients undergoing trans-sphenoidal pituitary surgery were usually discharged after 72 h. Those with intraoperative cerebrospinal fluid (CSF) leaks needed bed rest and intravenous antibiotics for 5–7 days. While in the hospital their serum sodium levels were closely monitored. After discharge from the hospital all of them were asked to come back for serum sodium testing on the seventh postoperative day. Patients with hyponatremia were readmitted and treated with 0.9% intravenous saline, intravenous hydrocortisone 50 mg 6 hourly, and oral salt 12 g over 24 h; those with serum sodium levels <125 mmol/L were treated with 3% intravenous saline.
We looked at the following risk factors in the patients who developed delayed hyponatremia: age, gender, intra-operative CSF leak, extent of resection (EOR), and post-operative DI [Table 2] and [Table 3].
Logistic regression analysis was used to identify the risk factors in patients who developed hyponatremia. Continuous variables were presented as means with standard deviations (SD). Categorical variables were presented as number and percent. Categorical variables were grouped as follows: gender, CSF leak, age (≤55 and >55 years), and EOR [gross total resection (GTR) and partial resection]. All potential risk factors were included in the logistic regression analysis. The effect modification of age and CSF leak was modeled. Hosmer and Lemeshow goodness-of-fit statistics was used to assess the fit of the model.
Fifty-eight patients (26%) developed delayed hyponatremia. Ten were on treatment for hypothyroidism and twenty six were on treatment for pre-existing hypocortisolism. All patients had macroadenomas; the average tumor size was 3.8 cm (range 1.7-6.6 cm) (Hardy's grade C/D/E). Gross total resection was achieved in 70% of patients. Sixteen patients (7%) with intraoperative CSF leak needed bed rest and intravenous antibiotics for 5–7 days. All the others were discharged after 72 h.
Delayed hyponatremia occurred mostly (43.1%) on the seventh post-operative day. The majority of our patients (81%) remained asymptomatic [Figure 1].
Severity of hyponatremia
Of the 58 patients who developed delayed hyponatremia, 25 had severe hyponatremia (sodium <125 mmol/L), 12 had moderate hyponatremia (sodium 125–129 mmol/L), and 21 had mild hyponatremia (sodium 130–134 mmol/L). The lowest serum sodium was 102 mmol/L. Urine spot sodium levels ranged from 8-243 mmol/L (median 170 mmol/L). The urine spot sodium levels did not correlate with the degree of hyponatremia. Twelve patients developed delayed hyponatremia while on oral prednislone.
The majority of our patients (81%) were asymptomatic. Eleven patients presented with vomiting (5), seizures (3), weakness (1), paralytic ileus (1), and fever (1). One patient developed status epilepticus. Serum sodium levels were lower in all patients with symptomatic hyponatremia as compared to patients with asymptomatic hyponatremia [mean 117.7 mmol/L (SD ± 2.1), median 113 mmol/L (range: 102–128) vs. mean 123 mmol/L (SD ± 1.1), median 131 mmol/L (range: 109–134)] (P < 0.01).
There was a higher incidence of delayed hyponatremia in those who developed post-operative DI as compared to those who did not develop post-operative DI (24% vs. 19.5%). Four patients developed the classical “Triple-phase response” (DI–SIADH–DI) due to recurrence of DI after the hyponatremia got corrected.
Fourteen patients had hypocortisolemia, ten had cortisol levels <5 μg/dL (severe hypocortisolism), and four had cortisol levels ranging from 5-10 μg/dL which was inappropriate for the degree of hyponatremia. The syndrome of inappropriate antidiuretic hormone (SIADH) was diagnosed in 3 patients who had hyponatremia and normal central venous pressure (CVP); these patients were treated with fluid restriction and oral salt. For patients who do not have CVP measurements we do not know the exact cause for the hyponatremia, it could have been either SIADH or cerebral salt wasting (CSW).
All patients received 12 g oral salt over 24 h in addition to 0.9% intravenous saline (50–100 mL/h). Those with serum sodium <125 mmol/L were treated with 3% intravenous saline. Most patients (92%) also received intravenous hydrocortisone 50 mg 6 hourly until their serum sodium levels normalized. One patient received oral prednisolone 10 mg daily instead of intravenous hydrocortisone. Two of the non-responders were treated with oral fludrocortisone 100 μg daily.
Ten patients with borderline and low thyroid functions after surgery (FTC 0.66-0.89) were treated with thyroxin.
Hyponatremia in most patients (57%) resolved within 48 h. Some took 3–7 days to respond [Figure 2].
As compared with females, males had 2.9 times higher odds (CI 1.4, 6.2) for developing delayed hyponatremia (P = 0.002). Similarly, those with intra-operative CSF leaks had 2.5 times higher odds for developing delayed hyponatremia (P = 0.003). Factors such as age, pre-operative cortisol levels, EOR, and post-operative DI did not correlate with the occurrence of delayed hyponatremia [Table 2] and [Table 3]. Patients who maintained their mean serum sodium levels >138 mmol/L in the immediate post-operative period (day 1- day 3) were unlikely to develop delayed hyponatremia (sensitivity 55.2% and specificity 83.9%), positive predictive value, 63.2% (CI 48, 76.7%), and negative predictive value, 78.8% (CI 70.6, 85.5%) [Table 4] and [Figure 3].
Incidence and temporal profile
Oertel et al. reported that delayed hyponatremia is the most common cause for readmission to hospital following pituitary surgery. The incidence of this complication ranges from 5% to 23%. Symptomatic hyponatremia has been reported to occur in 17%–78% of patients following pituitary surgery ,,,,, [Table 1].
The majority of our patients (81%) were asymptomatic. The peak occurrence of delayed hyponatremia was noted on the seventh postoperative day, this was similar to what was observed by other investigators , [Figure 1] and [Table 1]. Sata et al. mentioned that delayed hyponatremia can occur even as late as 9.5 ± 2.4 days after surgery.
Zada et al. mentioned that patients who developed postoperative DI had a 48% increased risk of later developing hyponatremia. This was seen in 24% of our patients.
Lee et al. found that gender, tumor type, and tumor size do not correlate with the occurrence of delayed hyponatremia. The only risk factor they could identify was age >50 years. Cote et al. in a systematic review of reports of symptomatic delayed hyponatremia found that age, gender, tumor size, rate of decline of blood sodium, and Cushing's disease are potential predictors for the occurrence of delayed hyponatremia. More recently, Krogh et al. mentioned that the risk of hyponatremia one week after surgery was higher in those with tumors abutting the optic chiasm and in those with declining sodium levels on the first post-operative day. In our study, male sex and intra-operative CSF leak were associated with an increased risk of developing delayed hyponatremia [Table 2] and [Table 3]. We found those who maintained mean serum sodium levels >138 mmol/L from day1- day3, were unlikely to develop delayed hyponatremia [Figure 4].
Delayed hyponatremia that occurs after the third post-operative day could be the result of either hypocortisolism, SIADH or CSW. Some believe that SIADH is the most common cause for delayed hyponatremia.,,, Sata et al. also demonstrated inappropriately high arginine vasopressin (AVP) levels in patients with delayed hyponatremia. More recent reports mention CSW as a contributing factor [Table 1].,,, In our study, 14 patients had hypocortisolism and 3 had SIADH. In the remaining patients the etiology remains unknown as we do not have their CVP measurements.
Our patients were treated with oral salt, intravenous saline, and intravenous hydrocortisone. Two patients in addition required oral fludrocortisone. Others have used different strategies to correct hyponatremia, Kelly et al. used intravenous urea, Richard et al. and Ferrante et al. found that a single dose dose of tolvaptan 15 mg (a competitive vasopressin receptor 2 antagonist) can quickly revert hyponatremia in these patients. Insertion of a central line and documenting CVP for all patients would have helped us differentiate SIADH and CSW.
Most patients in our study (57%) normalized serum sodium within 48 h [Figure 2]. Sata's  patients took 3.8 ± 1.7 days and Lee's  patients took 5 days to normalize serum sodium levels. More recently Richard  and Ferrante  report that their patients normalized serum sodium levels within 24 h.
Suggested post-operative screening protocol
Since the peak occurrence of delayed hyponatremia is on the seventh post-operative day, we recommend that all patients undergoing pituitary surgery come back for serum sodium testing on day 7 following surgery. This will help us identify all the asymptomatic cases before they develop symptoms.
Delayed hyponatremia occurred in 26% of our patients after surgery for nonfunctioning pituitary macroadenomas. This phenomenon tends to peak on the seventh post-operative day. Risk factors include the male gender, intra-operative CSF leak, and mean sodium levels <138 mmol/L in the first 72 h. Nausea and vomiting were the most common symptoms. Most patients remain asymptomatic and are identified on routine testing. Hypocortisolism contributes to 24%; SIADH and CSW contribute to the others. It is important to recognize this condition and initiate treatment early as serious complications can be avoided. Most patients recover within 48 h. It is important for all patients to come back for serum sodium testing on the seventh post-operative day.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4]