Hypocalcemia after total thyroidectomy may be transient (resolving within 6 months after surgery) or permanent (lasting more than 6 months after surgery). Transient hypocalcemia is the most common complication of total thyroidectomy and usually occurs between 24-48 hours (1). The incidence varies between 0.3% and 75% (2-4).
Postoperative hypocalcemia is caused by alkalosis secondary to hypoparathyroidism, hungry bone syndrome, hemodilution and hyperventilation caused by postoperative pain. The primary cause of transient hypocalcemia is hypoparathyroidism caused by intraoperative destruction of one or more parathyroid glands, iatrogenic parathyroid gland removal or nutritional disturbance (3,5-12).
The classic signs of hypocalcemia are associated with neuromuscular excitability. The clinical manifestations of postoperative hypocalcemia depend on the rate of onset and the degree of hypocalcemia. Symptoms may range from mild hypocalcemia with numbness around the fingers, feet and mouth to carpopedal spasm and tetany. The onset of symptoms may last up to 48-72 hours.
Transient hypocalcemia usually responds to calcium replacement therapy for a few days or weeks (13). It is important to predict which patients will develop hypocalcemia in order to start treatment early or to discharge the patient on a daily basis. In many studies in the literature, various factors have been evaluated to predict hypocalcemia developing after total thyroidectomy (1,8-10,13-26). However, a single factor that can definitely predict the development of hypocalcemia alone has not been demonstrated (2).
In this study, we aimed to create a risk scoring system with a combination of more than one factor to predict transient hypocalcemia after total thyroidectomy by considering these factors that may be related to hypocalcemia in the literature.
Materials and Methods
Biochemical and clinical data of 155 patients who underwent total thyroidectomy by a single surgeon between 2012 and 2013 were retrospectively evaluated.
Age, surgical intervention for recurrence, preoperative calcium, phosphorus, magnesium and alkaline phosphatase levels, preoperative hyperthyroidism status, number of parathyroid glands seen peroperatively, number of parathyroid glands transplanted, number of parathyroid glands in the pathologic specimen, postoperative 12th hour calcium, parathormone and magnesium levels were evaluated.
All evaluated parameters were assigned a score for the development of hypocalcemia (Table 1). Hypocalcemia was defined as a serum total calcium (Ca) level below 8 mg/dl. Transient hypocalcemia was defined as hypocalcemia that resolved within 6 months following total thyroidectomy. ''Multinomial Logistic Regression analysis'', ''Chi-square test'', ''Fisher's Exact Test'', ''Mann-Whitney U Test'' were used for statistical evaluation. p<0.05 was considered significant.
Findings
The distribution of the factors affecting hypocalcemia according to the patients and their statistical evaluation are shown in Table 2. Transient hypocalcemia developed in 22 of 155 patients (14.2%) in the study. In pairwise comparison, the presence of parathyroid gland in the specimen (p<0.01), early postoperative serum calcium level (p=0.0001), postoperative serum magnesium level (p=0.0001), and low postoperative serum parathormone level (p=0.0001) were statistically significant (Table 2).
Although the development of hypocalcemia could be explained by the parameters evaluated by logistic regression analysis with a rate of 63.2%, none of the factors were determined as independent risk factors. Since the scoring was not affected by gender (p=0.656), gender was excluded from the evaluation. In the scoring, a score of 3, 4, 5 and above was found to be a statistically significant independent indicator that could predict hypocalcemia (p<0.001). The sensitivity, specificity, positive predictive value, negative predictive value and diagnostic values according to the number of scores are given in Table 3.
Discussion
In the current literature, studies with biochemical values related to postoperative hypocalcemia generally focus on preop, perop and early postoperative serum calcium and parathormone level measurements. Studies have evaluated whether Ca and PTH levels can be used as predictive factors for hypocalcemia (18-20,27-32); however, the recommendations regarding the use of these factors in predicting hypocalcemia in all these studies do not agree with each other and there is no specific consensus or protocol (33). In different studies, the sensitivity of Ca measured on postoperative day 1 in predicting hypocalcemia has been reported in a wide range of 19-91% (15,34,35).
Pfleiderer et al. reported that a serum calcium level below 7.6 mg/dL on postoperative day 1 had a sensitivity of 95% in predicting clinical hypocalcemia (23). In addition, a greater decrease in Ca in the postoperative period compared to the preoperative period has also been found to be associated with transient hypocalcemia (36,37).
The positive predictive value of the presence of a positive Ca curve with an increase in the second value compared to the first value of two consecutive Ca values measured within the first 24 hours postoperatively in excluding transient hypocalcemia has been reported to be 86-100% (15,23,24). The positive predictive value of the negative Ca curve within 24 hours postoperatively in predicting transient hypocalcemia is significantly lower than the positive Ca curve and has been reported to be 20-46% (32,38,39).
In various studies, sensitivity rates for predicting hypocalcemia at peroperative and postoperative PTH values below the threshold value of 6-35 pg/mL have been reported in a wide range between 59-100% (15). Studies have reported 70-100% sensitivity and 70-99% specificity as predictive factors for postoperative 38-88% decreases compared to preoperative values (15).
Among these studies, Noordzij JP et al. (32) reported that a 65% decrease in the PTH value measured between 1-6 hours postoperatively compared to the preoperative value had 96.4% sensitivity and 91.4% specificity in predicting postoperative hypocalcemia in a meta-analysis of 457 patients from nine studies.
(34) reported that a PTH value below 15 pg/ml on postoperative day 1 had a sensitivity of 97.7%, a specificity of 82.6%, a positive predictive value of 65.6% and a negative predictive value of 99.1% in predicting hypocalcemia. When this value was combined with a Ca value below 1.9 mmol/L on postoperative day 2, the sensitivity and negative predictive value were reported to be 96.3%, 96.1%, and 86%, respectively, with no significant change.
(25) reported that the positive predictive value and specificity to exclude hypocalcemia were both 100% for values above 2.14 mmol/L Ca and 28 pg/mL PTH in the postoperative 6-12th hours. In the related study, 70-100% sensitivity and 70-99% specificity were reported as predictive factors for 38-88% postoperative decreases compared to preoperative values (15).
Although some studies have reported that preoperative vitamin D deficiency is associated with postoperative transient hypocalcemia (39-41), this relationship has not been demonstrated in other studies (42,43). In one study, it was reported that a postoperative Mg level below 0.7 mmol/L was associated with transient hypocalcemia independent of Ca+2 level (44).
In another study, although postoperative Mg level decreased, its role on postoperative hypocalcemia was not found (22). Phosphorus elevation before the 24th postoperative hour has also been reported to be a factor that may predict hypocalcemia in patients without vitamin D deficiency (45).
In addition to biochemical predictive factors for postoperative hypocalcemia, some clinical and surgical risk factors have also been reported. Among the patient-related factors, age and gender are among the most commonly evaluated factors. In a study by Tolone et al.(16), the risk of postoperative hypocalcemia was reported to be 20 times higher in patients over 50 years of age and accordingly, age was found to be significantly associated with postoperative hypocalcemia.
However, in a large retrospective study of 119,567 patients, Baldassarre et al. (46) reported that postoperative hypocalcemia developed less frequently in patients aged 45-84 years than in patients older and younger than the relevant age group. However, in this intermediate age group, the risk of hypocalcemia decreased by 1% with each increase in age and age had a protective effect.
In the meta-analysis conducted on this subject, no significant difference was found in terms of mean age between patients with and without transient hypocalcemia. In the same meta-analytic study, the incidence of hypocalcemia was found to be significantly higher in female gender (15). Female gender has been identified as a risk factor for hypocalcemia in univariate analysis or multivariate analysis in many studies (33,47-49).
In some studies, selective autotransplantation was found to be associated with transient hypocalcemia (23,34,36,48). Bergenfelz et al. (48) reported that this relationship was independent of the extent of thyroidectomy and neck dissection. Although there are studies reporting that incidental parathyroidectomy has no effect on postoperative hypocalcemia (40), many studies have found a higher incidence of transient hypocalcemia in patients with one or more incidental parathyroidectomies (26,50-53).
In addition to these features, reoperation for bleeding (48,54), postoperative wound infection (48), long operation time (54), performing the operation in university hospitals and hospitals with low surgical volume (54), and the extent of surgery (47,49,55) were found to have a higher incidence of transient hypocalcemia in multivariate analysis in different studies performed for thyroidectomy, Bilateral truncal ligation of the inferior thyroid artery near the carotid artery (49), thyroid cancer with macroscopic extrathyroidal extension (53), bilateral central neck dissection (22), recurrent goiter (47,49), Graves' disease (17), and weight of thyroid tissue removed (54) are reported as surgical and patient-related risk factors.
(2) defined a scoring system by evaluating 8 of these factors in their prospective study. They included age, hyperthyroidism, preoperative calcium and vitamin D levels, postoperative 12th hour calcium, postoperative 8th hour PTH level, peroperative parathyroid gland preservation status and nodule diameter as parameters.
According to the results of the study, when the hypocalcemia risk score was > 3, it showed 91% sensitivity, 84% specificity, 71% positive predictive value, and 95% negative predictive value; whereas with a score ≥ 4, 100% specificity and 100% positive predictive value were achieved, but sensitivity and negative predictive value decreased to 55% and 83%, respectively (2).
Researchers have stated that scores above 3 can be used as a predictive factor for the patients to be discharged from the hospital in 24 hours. In our study, 14 factors reported to affect hypocalcemia in the literature were examined. In our study, although sensitivity was 100% when the score was 3, specificity was 47%, positive predictive value was 24%, negative predictive value was 0% and diagnostic value was 54%. Although it is thought that these values can detect all hypocalcemic patients, only 25% of patients with a score above 3 will be hypocalcemic and 75% will be normocalcemic.
The positive predictive value, that is, the probability of hypocalcemia in patients with that score, increases to 36% when the score increases to 4 and 54% when the score increases to 5. When the score is 4 and 5, the specificity increases to 77% and 92%, the diagnostic value increases to 77% and 87%, respectively, while the negative predictive value increases only by 5% and 7%. When we start treatment by predicting that hypocalcemia may develop in patients according to these scores, unnecessary treatment is given to 75% of those with a score above 3, 64% of those with a score above 4, and approximately 50% of those with a score above 5. Or, when we do not give treatment and observe these patients by hospitalizing them, the patient who will not develop hypocalcemia at the specified rates will be hospitalized for a long time.
When the negative predictive value, that is, when the score is below the specified value, the probability of the patient being normocalcemic is extremely low. In other words, this scoring system is inadequate to predict patients who are normocalcemic and can be safely discharged. This system, which includes many factors and requires knowledge of more tests and features than a single test, does not seem to be a system that can make an additional contribution to the prediction values reported for postoperative calcium and PTH values discussed above.
The fact that these ratios, which express the degree of accuracy of the tests performed in our study, show differences with the other study may be related to the different characteristics evaluated as well as the different patient population. The retrospective nature of the study is the weakness of the study.
Postoperative hypocalcemia is multifactorial. In our study, only 63% of the evaluated features could explain postoperative hypocalcemia and none of them was found to be an independent risk factor.
In conclusion, although the presence of 4,5,6 scores are independent indicators in predicting postoperative hypocalcemia, they cannot be recommended as a definitive scoring system due to low negative and positive predictive values. Since hypocalcemia is significantly higher in patients with low early postoperative calcium, magnesium and parathormone levels, these patients should be carefully monitored for the development of hypocalcemia. An effective scoring system with high predictive power can be established with larger prospective studies in the future.
