Introduction

Total thyroidectomy is the standard surgical procedure for the management of thyroid disease, such as bilateral nodular thyroid, Grave's disease, goiter and cancer. One of the most common complications is hypocalcemia, the reported incidence varies from 1 to 50% [1, 2]. Hypocalcemia after thyroid surgery is linked to a diminished circulating parathormone (PTH) as a result of injury to the parathyroid glands by direct injury or injury to the vascular supplies. Clinical symptoms are perioral numbness, cramps, paresthesias of the upper extremity digits, Chovstek’s sign, and convulsions. These main symptoms occur between 24 and 48 h postoperatively and are the primary reason of prolonged hospital stay, once the risks of cervical hematoma and airway obstruction have decreased [3, 4]. Measurement of PTH may reduce the hospital stay by predicting which patients will require calcium monitoring and postoperative supplementation. Numerous studies [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20] reported the advantages of measurement of PTH. But the time of collection of the blood sample and the cut-off value of PTH were not well defined and varied according to different authors. Some authors reported their results in terms of absolute value of PTH (ng/L). Other authors reported their results in terms of decrease in the slope (ng/L) and decline (%) between two times of blood sample collection, this approach required the collection of several samples. For the comparison of the different analysis methods, we studied the three possibilities. The main purpose of our study was to investigate which time of collection of the blood sample and which cut-off value of serum intact parathormone (iPTH) levels could predict postoperative hypocalcemia.

Methods

We conducted a prospective study in patients who underwent total thyroidectomy in our University Hospital between august 2012 and august 2013. Patients with a history of thyroid or neck surgery, concomitant parathyroid disorder, renal failure, calcium or vitamin D supplementation were excluded. Local ethic committee approved the study. All patients gave written consent for the study. All the surgical procedures, extra-capsular total thyroidectomy, were performed under general anesthesia, by the same experienced surgeon (24 years’ experience). Operative time and final thyroid pathology were recorded. To preserve maximum parathyroid glands, all vascular ligatures used clips placed as close as possible to the thyroid gland. If signs of suffering parathyroid glands were observed or if vascular supply of the parathyroid was injured, the gland was placed in the right sterno-cleido-mastoid muscle at the end of the procedure. Back from the surgical unit, patients were monitored in hospitalization unit for two nights. Particular attention was paid to the development of early clinical signs of hypocalcemia. Patients with symptomatic or asymptomatic hypocalcemia (< 1.8 mmol/L) were started on one of the following regimens on the basis of calcium level and clinical symptoms: intravenous calcium, oral calcium, oral vitamin D, or a combination of the three regimens. The patient's discharge was permitted on the second postoperative day providing there were no complications and a normal calcium level. If necessary, patients were discharged on calcium supplementation. All patients were reviewed with a biochemical control of calcium levels one week and one month after the surgery.

Blood samples for iPTH, serum calcium (Ca) and serum albumin (Alb) were drawn from a peripheral intravenous line, preoperatively (H0), immediately after removal of the thyroid gland (Hdrop), 6 h (H6) and one day (D1) after the end of surgery. Another, including calcium and albumin, was performed on the second postoperative day (D2). The blood samples for iPTH had to be handled according to a specific protocol to be analyzed. They were placed into ice and carried at + 4 °C from the operating room to the laboratory. They were centrifuged for 3 min at the same temperature. Serum was then frozen at − 20 °C until biochemically analyzed. Twice a week, biochemical analyses of iPTH were performed by an automated immuno-chemiluminescence assay, intact PTH kit on ISYS® (IDS, UK). Several times a day, for calcium, a spectrophotometric method was performed in the hospital laboratory on ADVIA® 1800 PLC (Siemens, Germany). The normal values for different assays ranged from 2.10 to 2.65 mmol/L for calcium and from 10 to 55 ng/L for iPTH. To minimize the effect of hemo-dilution, the albumin-adjusted total serum calcium (Cac) levels were calculated. It was defined by Cac (mmol/L) = [40 − Alb (g/L)]/40 + Ca (mmol/L). Decrease in the slope (ng/L) of iPTH and decline (%) of iPTH, between two times, were, respectively, defined by ΔiPTH = iPTHHo − iPTHH6 and %ΔiPTH = 100(iPTHH6/iPTHHo).

Symptoms and level of hypocalcemia were recorded. Hypocalcemia was defined as total calcium values corrected for albumin ≤ 2.10 mmol/L, and/or with presence of hypocalcemia clinical signs during hospitalization. Patients were divided into two groups: group 1 consisted of patients with hypocalcemia on the second postoperative day (D2), and group 2 consisted of patient with normocalcemia on D2.

Patients who had hypocalcemia were compared with normocalcemic patients. The area under the ROC curve (AUC) was used to determine thresholds and predictability of iPTH for hypocalcemia. The different absolute values of iPTH (ng/L) at each time of blood sample collection, the decrease in the slope (ng/L) and the decline (%) between preoperative and postoperative time were analyzed. For each threshold, sensitivity (Se), specificity (Sp), and positive and negative predictive values (PPV and NPV, respectively) were estimated with their confidence interval of 95%. To improve clinical utility, Youden’s index was used to calculate the best cut-off value for predicting hypocalcemia. For the comparison between the two groups, Kruskal–Wallis test was used for surgical indication, t test was used for comparison of age and operative time, Chi2 test was used for sex. A probability value p < 0.05 was considered significant. Data were analyzed using SPSS version 17.0 statistical software (SPSS, Inc, Chicago, Illinois).

Results

Between August 2012 and August 2013, 108 consecutive patients were eligible to the study. Seven patients were excluded (Preoperative blood calcium was not available for 4 patients and 3 patients were hypocalcemic before the surgery). One-hundred-and-one patients were finally included in the study. Among them, 18 were men (17.8%) and 83 were women (82.2%) with a median age of 50.2 years (SD = 15.1; range 18–83) (Table 1). The indication for surgery was Graves’s disease for 33 patients (32.7%), goiter in 50 patients (49.5%), suspicious nodule for 14 patients (13.9%), and thyroiditis in 4 patients (3.9%). Pathological examination revealed 8 papillary adenocarcinoma, 2 medullar carcinoma and 15 papillary micro carcinomas. In this subgroup, the frequency of hypocalcemia was the same. The mean duration of surgery was 64.78 min (SD = 18.22).

Table 1 Patients characteristics (n = 101)
Full size table

At each period of blood sample collection, some results could not be obtained due to hemolysis of the samples or non-fulfillment of the transportation conditions. The biochemical assays of iPTH and serum calcium level are summarized in Table 2.

Table 2 Biochemical assays of calcium (mmol/L) and iPTH serum level (ng/L) for all patients at different times of blood sample collection (Ho, Hdrop, H6, D1, and D2)
Full size table

Two groups have been created, group 1 included the hypocalcemic patients on the second postoperative day and group 2 included the normocalcemic patients. Patients of both groups did not differ in age (p = 0.332), sex (p = 0.059), operative time (p = 0.085) (Table 3), surgical indication (p = 0.172) and serum level of preoperative iPTH (54.60 ng/L for hypocalcemic patients and 44.90 ng/L for normocalcemic patient; p = 0.575). However, serum levels of preoperative calcium were lower in the hypocalcemia group than in the normocalcemia group (2.22 mmol/L [2.10; 2.44] versus 2.26 mmol/L [2.10; 2.42]; p = 0.035).

Table 3 Comparison of variables between the hypocalcemic and normocalcemic groups, and description of the patients who were hypocalcemic after one month of follow-up
Full size table

Postoperatively, statistical analysis demonstrated significant difference (p < 0.001) between the two groups with regard to all dates (iPTHdrop, iPTHH6, CacH6, iPTHD1, CacD1, CacD2) (Table 4).

Table 4 Comparison of median iPTH (ng/L) and calcium (mmol/L) between hypocalcemic and normocalcemic patients
Full size table

On the second postoperative day, 39 patients (38.6%) were hypocalcemic. According to the procedure of our department, 12 patients (11.9%) were symptomatic for hypocalcemia, all of them received postoperative calcium. Symptoms of hypocalcemia were paresthesias of the upper extremity digits (n = 8), perioral numbness (n = 5), and cramps (n = 2). Two (2%) were asymptomatic but received supplementation because of a postoperative concentration of calcium less than 1.80 mmol/L. The onset of symptoms occurred between 12 and 48 h after surgery. The collection of blood sample, one week after the surgery, did not highlight more hypocalcemia. One month after the surgery, only 8 patients remained hypocalcemic (7.9%). At 6 months, none of these 8 patients had hypocalcemia. Among patients with hypocalcemia, concentration of postoperative iPTH was not lower in this 8 patients, compared to the others, respectively: iPTHdrop 11.22 ng/L vs 12.39 ng/L p = 0.318, iPTHH6 5.89 ng/L vs 6.84 ng/L p = 0.350, iPTHD1: 7.59 ng/L vs 7.22 ng/L p = 0.652.

Patients who developed postoperative hypocalcemia had a significant lower postoperative concentration of postoperative iPTH (iPTHdrop, iPTHH6, iPTHD1) than normocalcemic patients (Table 4). ROC curves were obtained for each time of blood sample (Figs. 1, 2, and 3). The best area under the ROC curve was identified for iPTHH6 AUC = 0.857 [0.769; 0.944] (Table 5). Absolute iPTHH6 was more discriminating than other absolute values for early prediction of hypocalcemia on the second postoperative day. At this time, the better threshold (Table 6) obtained was iPTHH6 = 14.35 ng/L (Youden’s index = 0.623) with Se = 0.706 [0.523; 0.843]; Sp = 0.917 [0.809; 0.969]; PPV = 0.828 [0.635; 0.935] and NPV = 0.846 [0.731; 0.920]. We considered this value as the cut-off point. When the iPTHH6 was below 5 ng/L, the PPV was of 1.000 [0.717; 1.000]. When the iPTHH6 was above 74.8 ng/L, the NPV was of 1.000 [0.320; 1.000].

Fig. 1
figure 1

ROC curves of absolute values of iPTH correlated with the albumin­ adjusted serum calcium level on the second postoperative day. iPTH intact parathormone

Full size image
Fig. 2
figure 2

ROC curves of decrease in the slope (ng/L) of iPTH correlated with the albumin-adjusted serum calcium level on the second postoperative day. ΔiPTH decrease in the slope of intact parathormone between preoperative time and different times of blood sample collection

Full size image
Fig. 3
figure 3

ROC curves of decline (%) of iPTH correlated with the albumin­ adjusted serum calcium level on the second postoperative day. %ΔiPTH Decline of iPTH between preoperative time and different times of blood sample collection

Full size image
Table 5 Area under the iPTH ROC curves for different times of blood sample collection with their confidence interval at 95% [absolute value (ng/L), decrease in the slope (ng/L) and decline of iPTH (%)]
Full size table
Table 6 Summary of different absolute values iPTH thresholds identified and their Se/Sp/PPV/NPV associated with confidence intervals at 95%
Full size table

In terms of decline of iPTH, the best AUC was obtained for the decline between the preoperative time and six hours after the surgery (AUC = 0.856 [0.770; 0.953]). At this time, the cut-off value was 59.5% with Se = 0.85 [0.68; 0.94]; Sp = 0.82 [0.69; 0.90]; PPV = 0.74 [0.58; 0.86] and NPV = 0.90 [0.77; 0.96]. For the ROC curves of serum calcium level on the first postoperative day correlated with the serum calcium level on the second postoperative day, AUC = 0.921 [0.859; 0.982].

Discussion

To compare our study with previous (Table 7), iPTH was measured at different times already used in the literature as a predictor of postoperative hypocalcemia [5, 6, 10, 11]. A possible difference between studies is the duration of the surgery, in our study of about one hour, which is not always mentioned. Using the ROC curves, we were able to determine the best collection time of blood sample for iPTH to predict hypocalcemia on the second postoperative day. With the best area under the curve, iPTH, six hours after surgery, was predictive of hypocalcemia. It was in agreement with literature data. Noordzij et al., in an analysis of 9 observational studies [21], revealed that iPTHH6 was predictor of postoperative hypocalcemia (Se = 96.4%; Sp = 91.4%). Lombardi et al. [7] founded the iPTHH6 had the best predictive value. In our study, all the patients with iPTHH6 under 5 ng/L had a low level of calcium and needed a supplementation to prevent symptoms of hypocalcemia. Contrariwise, all the patients with iPTHH6 higher than 74.80 ng/L had not a low level of calcium on the second postoperative day. Six hours after the surgery, iPTH threshold of 14.35 ng/L was predictive of hypocalcemia post thyroidectomy total with Se = 0.706; Sp = 0.917; PPV = 0.828 and NPV = 0.846. In this prospective study, with the same experienced surgeon, we studied many parameters in this cohort. We showed that the use of iPTHH6 dosage would permit to take care of patients with high risk of hypocalcemia before onset of symptoms. It would reduce the hospitalization stay for the patients with low risk of hypocalcemia. A short stay in hospital can lead to health care savings without increasing the risk for the patient.

Table 7 Published trials about the use of iPTH levels to predict postoperative hypocalcemia
Full size table

Percentage of decline of iPTH between preoperative blood sample and six hours after the surgery had seemed to be less interesting even if AUC was similar. It did not allow us to establish a threshold with PPV = 100% nor NPV = 100%. As for iPTHH6, we confirmed the literature data for the intraoperative iPTH: it was less sensitive and specific than other times [23]. As Payne et al. [6], we did not find iPTHD1 was most sensitive in prediction than iPTHH6. Lindbolm et al. [5], showed that iPTH on the first postoperative day was not more sensitive than albumin-adjusted serum calcium level. We confirmed that literature data, the AUC of calcemia on the first operative day was better than AUC of iPTHH6 to predict hypocalcemia on the second postoperative day. Despite this lower predictability, the assay of iPTH could allow an earlier supplementation. Moreover, iPTH measuring was more expansive than blood sample analysis for albumin-adjusted serum calcium level (16.20 euros versus 5.88 euros, respectively). Above all, transportation conditions for iPTH measurement were more stringent causing more failure. In this study, patients were not supplemented earlier because 6 h blood results were not obtained immediately.

Any injury leading to impairment in PTH secretion leads to an immediate decline in iPTH serum levels. Our data showed a decline in PTH after the surgery (Hdrop, H6, and D1), we confirmed the susceptibility of the parathyroid glands to surgical trauma. Our hypocalcemia rate of 38.6%, on the second postoperative day, was in agreement with literature data (16 to 52%) [5,6,7,8,9,10,11,12,13]. Unlike other studies [22], we did not reveal any influence of the surgical indication on the development of hypocalcemia. Factors, such as age, sex, operative time and serum level of preoperative iPTH, were not predictors for development of hypocalcemia. However, preoperative serum calcium levels were lower in the hypocalcemia group than in the normocalcemia group [23]. All the patients with normal albumin-adjusted serum calcium level on the second postoperative day, remained normocalcemic one week later. One month after the surgery, the hypocalcemia rate was 8%. There were no permanent hypocalcemias. The auto-transplantation parathyroid rate of 3.0% was lower than other studies [9]. There was no significant difference of postoperative calcemia between auto-transplanted patients and the others, unlike other study [9]. This lack of significant difference may be due to the small number of auto-transplanted patients.

Unfortunately, we did not have more than one-month follow-up, except for the 8 patients with hypocalcemia at one month. Moreover, biochemical assay of 25 hydroxyvitamin D before surgery was not carried out.

Conclusion

The determination of iPTH six hours after end of the surgical procedure was predictive of postoperative hypocalcemia, in terms of absolute value and decline, with good sensitivity and specificity. However, we were not able to show significant difference between the absolute value of iPTHH6 and the decline of ΔiPTHH0H6.

A new protocol was developed:

  • IPTHH6 < 14.35 ng/L: oral vitamin D (Calcitriol—1,25-dihydroxycholecalciferol) 0.5 µg / 8 h and calcium 500 mg/8 h; control of serum calcium level 12 h after start of treatment or sooner if symptoms persist, treatment adjustment and release of the patient according to the serum level calcium;

  • IPTHH6 > 74.8 ng/L: release of the patient on the first postoperative day without supplementation and without control of serum level calcium;

  • iPTHH6 = 14.35–74.8 ng/L =  > control of serum calcium level at 12 o’clock on the first postoperative day or sooner if symptoms of hypocalcemia, treatment adjustment and release of the patient according to the serum level calcium. The purpose of this new protocol of treatment was to reduce the hospitalization for patients at low risk of hypocalcemia and for asymptomatic patients after earlier supplementation.