- Email: [email protected]
Parathyroid hormone early percent change: an individualized approach to predict postthyroidectomy hypocalcemia
Available online at www.sciencedirect.com
American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216 – 220 www.elsevier.com/locate/amjoto
Parathyroid hormone early percent change: an individualized approach to predict postthyroidectomy hypocalcemia☆ D. Brandon Chapman, MD a , Christopher C. French, MD a , Xiaoyan Leng, MD b , J. Dale Browne, MD a , Joshua D. Waltonen, MD a , Christopher A. Sullivan, MD a,⁎ a
Department of Otolaryngology/Head and Neck Surgery, Wake Forest University School of Medicine, Winston Salem, NC, USA b Department of Biostatistical Sciences, Wake Forest University School of Medicine, Winston Salem, NC, USA Received 21 March 2011
Abstract
Purpose: The purpose of this study is to evaluate a percent change model of postoperative parathyroid hormone level in thyroidectomy patients as a predictor of hypocalcemia. Materials and methods: Chart review was completed on patients who had undergone total or completion thyroidectomy over a 22-month period in our department. Only those patients with a preoperative ionized calcium and parathyroid hormone (PTH) level and at least 1 postoperative result were included. Ionized calcium levels served as an internal control. The Student t test was used to compare PTH level between the normocalcemic and hypocalcemic groups at each time point. Logistic regression analysis was used to predict hypocalcemia based on the diagnostic criteria. Receiver operator curves were used to maximize sensitivity. Results: Fifty-two patients met the inclusion criteria during the study period. A total of 22 patients (42%) experienced hypocalcemia. We were unable to maximize both sensitivity and specificity at the same time point. When comparing preoperative to 6-hour postoperative PTH percent change, patients with a greater than 44% decrease are likely to have hypocalcemia, with a sensitivity of 100%. Likewise, in those patients without a greater than 44% decrease at 6 hours, early discharge can be considered safe, given the negative predictive value of 100%. Conclusion: In our series, patients with a greater than 44% PTH decrease from preoperative to 6hour postoperative are very likely to develop hypocalcemia. We would propose that these patients need further inpatient monitoring to progress to safe discharge. Likewise, patients with a less than 44% decrease at the 6-hour time point are unlikely to develop hypocalcemia and may be considered safe for discharge. © 2012 Elsevier Inc. All rights reserved.
1. Introduction Hypocalcemia is the most common etiology for prolonged hospitalization in patients undergoing total thyroidectomy [1]. Traditional methods for detecting hypocalcemia in the postoperative period include frequent monitoring of total and ionized calcium (ICa) levels as well as close clinical ☆ The authors have no conflicts of interest or external funding sources to disclose. ⁎ Corresponding author. Department of Otolaryngology, 4th Floor Watlington Hall, One Medical Center Blvd, Winston Salem NC, 27157, USA. E-mail address: [email protected] (C.A. Sullivan).
0196-0709/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.amjoto.2011.06.004
monitoring for symptoms of hypocalcemia. These traditional methods can take 24 hours or longer to reveal hypocalcemia, which can prolong hospitalization. A clinical laboratory method for early prediction of postoperative hypocalcemia could, therefore, facilitate earlier discharge (≤24 hours) and improve outcomes after total thyroidectomy. The risks of postoperative hematoma and airway compromise are highest during the first 24 hours after thyroidectomy. However, symptomatic hypocalcemia may not manifest itself during the first 24 hours [2]. With the advent of pay-for-performance measures, many endocrine surgeons are interested in early predication of hypocalcemia, both to prevent serious complications and to begin early treatment in an effort to hasten hospital discharge.
D.B. Chapman et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216–220
Over the past 20 years, clinical research has focused on determining an efficient method for monitoring and predicting postoperative hypocalcemia. Initial efforts focused on early perioperative serum calcium slopes; however, these ultimately proved to have limited negative predictive values (NPVs) [3,4]. In recent years, multiple retrospective studies have emerged, which support the use of postoperative serum parathyroid hormone (PTH) levels as accurate predictors of hypocalcemia in postoperative thyroidectomy patients. A recent meta-analysis examined the pooled data from 9 individual studies examining postoperative PTH levels as predictors of hypocalcemia. This demonstrated that PTH levels at 6 hours postoperatively had a sensitivity of 96.4% and a specificity of 91.4% for detecting postoperative hypocalcemia [5]. Several studies have attempted to correlate PTH levels with clinical outcome while still using calcium levels and clinical picture as the criterion standard. Vescan et al [6] performed a prospective study of 201 patients undergoing total thyroidectomy. Postoperative PTH levels were assessed 1 hour after the procedure and also the following morning. Standard calcium levels and symptom monitoring were used to determine disposition. Results suggested that patients with PTH levels greater than 1.6 pmol/L at 1 hour were likely to remain normocalcemic. Patients with PTH levels less than 1.1 pmol/L at 1 hour were more at risk for developing hypocalcemia. Scurry et al [7] also evaluated postoperative PTH levels in thyroidectomy patients. They determined that accurate predictors for the development of symptomatic hypocalcemia included a decrease in PTH by 75% as well as an absolute postoperative PTH of 7 pg/mL. Payne et al [8,9] have published considerable research on this topic as well. In several articles, they have demonstrated that a PTH greater than 28 ng/L (pg/mL) combined with a normal calcium level at 6 hours postthyroidectomy can be predictive of patients that will maintain normocalcemia. They accomplished a specificity of 100%, sensitivity of 67.9%, and 100% positive predictive value. Based on the aforementioned studies, we have been using patient PTH levels in the perioperative period to assist in decision making and discharge planning since January 2008. However, because of differing reports and variable conclusions in the literature, we have only been able to use these data as an adjunct to clinical decision making while still relying on the criterion standard of ICa levels to predict postoperative hypocalcemia. We have observed that most reports suggest using an absolute cutoff value as a predictor of potential hypocalcemia. We find this approach to be problematic due to the wide range of reporting units (picomolar, picogram per milliliter, and nanogram per liter) and the existence of inherent variability between the assay type/methodology and logistics at different clinical laboratories. Preoperative PTH levels might also be an important consideration, especially in patients with abnormal preoperative parathyroid function. As such, the overarching goal of our study was to examine the
217
use of a percent change model for serum PTH level to predict postoperative hypocalcemia after thyroidectomy.
2. Materials and methods 2.1. Study design After obtaining institutional review board approval, we retrospectively reviewed hospital and clinic records of patients who had undergone total or completion thyroidectomy in our department from January 1, 2008, to October 30, 2009. Primary surgeons included the 3 senior authors (JDB, CAS, and JDW). Ionized calcium levels were used for the internal control group. Intact PTH levels (PTH) were obtained using a standard commercial immunoassay, which measures 1.84 intact PTH molecule (institutional reference range, 12–72 pg/mL). Parathyroid hormone levels were evaluated at the following time points: preoperatively (any time from 30 days before the procedure to immediately before incision), immediately postoperative (exactly 10 minutes after passing off the surgical specimen), and 6 hours postoperatively. Ionized calcium levels were obtained at the following time points: preoperatively; immediately postoperative; and then 6, 12, 24, 36, and 48 hours postoperatively, followed by daily levels for the remainder of hospitalization. These values were rarely obtained after the 24-hour time point, unless the patient remained in the hospital secondary to high drain output, unresponsive or worsening hypocalcemia, or other medical comorbidities. Patients experiencing symptoms of hypocalcemia or with ICa less than 1.00 mmol/L (institutional reference range, 1.00–1.30 mmol/L) were started on oral calcium supplementation and calcitriol immediately. Patients with significant hypocalcemia (ICa b0.80 mmol/L, severe symptoms, or known cardiac conduction problems) were treated with intravenous calcium gluconate, in addition to initiation of oral calcium supplementation and calcitriol. 2.2. Inclusion criteria To be included, the patient required both preoperative and at least 1 set of postoperative PTH and ICa levels. Data reviewed included patient age, sex, date of surgery, primary surgeon, procedure performed, pathologic diagnosis, laboratory values (ICa and PTH), and symptoms. 2.3. Exclusion criteria Patients were excluded from the study if they were undergoing an extended procedure such as total laryngectomy or parathyroidectomy. They were also excluded if they were taking preoperative osteoporotic medications or prophylactic vitamin D and calcium supplementation. Patients with preexisting chronic renal disease, hyperparathyroidism, or known digestive disease were also excluded.
218
D.B. Chapman et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216–220
2.4. Outcome measures Our primary outcome measure was the development of hypocalcemia, as predicted by percent change in PTH level. Hypocalcemia was defined as either a fall in the patient's ICa level below 1.00 mmol/L or signs/symptoms of hypocalcemia during the postoperative period (up to 14 days postoperatively). Commonly accepted signs of hypocalcemia include the Chvostek sign, Trousseau sign, and cardiac arrhythmias. Common symptoms of hypocalcemia that were assessed include perioral paresthesias, distal digital paresthesias, tetany, and palpitations (as documented in the patient's daily progress notes or postoperative clinic notes). 2.5. Statistical analysis Patient demographics and basic characteristics were presented as mean/SD for continuous variables such as age. Patient demographics and basic characteristics were presented as percentages for categorical variables such as sex, procedure, pathologic diagnosis, signs/symptoms of hypocalcemia, and hypocalcemia proven by laboratory values. Mean PTH levels were calculated at each time point (preoperatively, immediately postoperatively, and 6hour postoperatively), and the 2-sided Student t test was performed to compare the normocalcemic and hypocalcemic groups. Logistic regression was used to predict hypocalcemia based on the diagnostic criteria. Sensitivity and specificity were then calculated based on different cutoff points from the logistic regression model. Receiver operator curves were plotted with sensitivity vs 1-specificity. The cutoff point with maximum sensitivity was chosen as optimal for this screening test. Diagnosis criteria included both absolute PTH change and percent change for the immediate postoperative time point and the 6-hours postoperative time point. Sex effect was also examined using logistic regression model.
3. Results There were a total of 320 thyroidectomy cases performed by the Wake Forest University Otolaryngology service, which were reviewed for inclusion criteria over a period of 20 months. A total of 52 (16% of all reviewed) patients met the inclusion criteria during the study period. The most common reasons for exclusion included a hemithyroidectomy, preoperative calcium supplementation, and lack of required laboratory values at the specified time points. Mean age of our study population was 50.2 years, and 69% were female. Papillary carcinoma was identified in 58% of patients. Most patients (94%) underwent total thyroidectomy (vs completion thyroidectomy). Central compartment neck dissection was included in the procedure in 13 (25%) of the patients. A total of 22 patients (42%) experienced hypocalcemia (based on symptoms or ICa level). The complete data set including patient age, sex, procedure performed, final
pathologic diagnosis, incidence of signs/symptoms of hypocalcemia, central neck compartment dissections, and incidence of hypocalcemia is displayed in Table 1. The mean preoperative PTH was 69.0 pg/mL, with no significant difference between the normocalcemic and hypocalcemic groups (P = .2317). However, a significant difference existed in mean PTH levels when comparing the normocalcemic and hypocalcemic groups during the immediate postoperative and 6-hour postoperative time points (P = .0144 and P = .0087, respectively). Mean values for each group are included in Table 2. Sex was not found to have effect on the incidence of hypocalcemia using either the absolute change or percent change models. We were unable to maximize both sensitivity and specificity at the same time point. Therefore, given that the main goal of this study was to establish an accurate screening test, we calculated a cutoff point that could independently maximize the sensitivity. Receiver operator curves were used to achieve this task. It was found that within the immediate postoperative period, results were too variable to be applicable to the study goals. The 6-hour time point appeared optimal for maximizing sensitivity. This analysis (Table 3) revealed that when comparing the percentage of change from preoperative to 6-hour postoperative PTH levels, patients with a greater than 44% decrease were more likely to have hypocalcemia (sensitivity, 100%) (Fig. 1). Likewise, patients having a
Table 1 Patient demographics and basic characteristics Study group
N = 52
Mean age in years (SD) Sex Male Female Procedure Total thyroidectomy Completion thyroidectomy Pathologic diagnosis Papillary carcinoma Multinodular goiter Hashimoto's thyroiditis Medullary carcinoma Follicular adenoma Hurthle cell adenoma Thyroglossal duct cyst Signs/symptoms of hypocalcemia Digital paresthesia Chvostek sign Palpitations Perioral paresthesia None Central neck compartment dissections Yes No Hypocalcemia (ICa b1.0, or Si/Sx) Yes No
50.2 (18.0) 16 (31%) 36 (69%) 49 (94.2%) 3 (5.8%) 30 (58%) 12 (23%) 5 (10%) 2 (4%) 1 (2%) 1 (2%) 1 (2%) 4 (8%) 1 (2%) 1 (2%) 0 (0%) 46 (88%) 13 (25%) 39 (75%) 22 (42%) 30 (58%)
D.B. Chapman et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216–220
219
Table 2 Mean differences in PTH at different time points
Preoperative Immediate postoperative 6-h postoperative
All patients
Normocalcemic
Hypocalcemic
P⁎
69.0 ± 39.9 27.7 ± 23.8
63.3 ± 41.5 34.7 ± 23.8
76.8 ± 37.2 18.5 ± 21.0
.2317 .0144
24.3 ± 23.3
32.2 ± 20.7
14.9 ± 23.1
.0087
Data presented as mean ± SD (picogram per milliliter). ⁎ P value compares normocalcemic mean PTH to hypocalcemic mean PTH.
less than 44% decrease at 6 hours could be considered safe for early discharge, given the NPV of 100%.
4. Discussion Traditional serum calcium monitoring is highly specific, yet this is also time consuming. Because of the physiologic time lag between laboratory draws and the manifestations of hypocalcemia, this method is intrinsically flawed and will always demonstrate low sensitivity during early serum calcium measurements. It was our goal to determine if early percent change in serum PTH levels could prove to be an early and accurate predictor of hypocalcemia. The ideal screening test maximizes sensitivity so that no patients are “missed.” Using the percent change model to accurately predict eventual hypocalcemia, this study demonstrated a maximum sensitivity when the change in PTH levels from preoperatively to 6-hour postoperatively decreased by 44% or greater. This can aid in early identification of patients who will likely develop hypocalcemia and, thus, will likely benefit from early supplementation. After beginning supplementation, we recommend traditional calcium testing because PTH levels would be skewed by exogenous calcium. Likewise, this study demonstrated an NPV of 100% when comparing the preoperative and 6-hour postoperative levels. This infers that patients not exceeding a 44% drop in PTH level at 6 hours will likely not develop hypocalcemia. We would propose that these patients can be safely discharged once the risk for other perioperative complications has passed. These patients should be given warnings for the
Table 3 Diagnosis criteria comparisons, maximizing sensitivity Test
Pre—6-h absolute change (pg/mL)
Pre–6-h postoperative % change
Cutoff value Sensitivity Specificity Positive predictive value NPV
−24 1 0.423 0.595
−44 1 0.538 0.647
1
1
Fig. 1. Six-hour percent change vs sensitivity.
signs and symptoms of hypocalcemia but would likely not need additional laboratory draws. Our optimization of sensitivity comes at the expense of specificity. Although the percent change model can be established as an excellent screening test, one must keep in mind that there will be a higher rate of false positives. Although inappropriate treatment with calcium and vitamin D analogs can lead to hypercalcemia and subsequent symptom, short-term “overtreatment” is, in the authors' minds, less risky than undertreatment. We would prefer to capture all potentially hypocalcemic patients, as we have seen more severe complications from undertreated patients than from overtreated patients. Many institutions [10], including our own, have in the past routinely treated patients with prophylactic oral calcium and vitamin D with a low incidence of complications from hypercalcemia. With that in mind, careful postoperative follow-up and rapid weaning of supplementation should be sought. Although there has been an abundance of retrospective studies examining the predictive powers of postoperative PTH levels, true outcome studies have been limited. Sabour et al [11] retrospectively demonstrated that the use of rapid post anesthesia care unit (PACU) PTH levels in decision making for supplementation can impact the overall rate of hypocalcemia in thyroidectomy patients. They discovered that use of immediate PTH levels in the PACU significantly reduced the proportion of patients experiencing both mild and severe hypocalcemia but did not completely prevent hypocalcemia in all patients. In this cohort, decisions were based on absolute values of PTH. A recent literature review by Grodski and Serpell [12] reviewed 27 series and concluded that a single PTH measurement drawn any time from 10 minutes postoperatively to several hours later will provide equally accurate results. This study is not without limitations. The retrospective nature of this study makes it difficult to ensure complete capture of the signs and symptoms of hypocalcemia. The low rate of patient capture due to missing laboratory values is also a reflection of the retrospective nature. In addition, there is little consensus in the literature as to what ICa levels should prompt intervention or what intervention should be undertaken [13]. Our definition of hypocalcemia, based on ICa less than 1.00
220
D.B. Chapman et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216–220
mmol/L, may capture patients who would not have developed signs or symptoms of hypocalcemia. Our preference, however, is to favor the capturing of false positives, rather than miss potential hypocalcemic patients. The retrospective nature of this study also limited our ability to ensure preoperative PTH and ICa laboratory evaluation, which significantly reduced our sample size due to exclusion criteria. Another limitation that has plagued many similar studies has been a difficulty in the classification of patients with regard to their preoperative risk factors, specifically related to their medication use. Many patients are on over-thecounter regimens including calcium and vitamin D, which is often difficult to capture in a retrospective series. This is further confounded by the increasing use of bisphosphonates for osteoporosis, which can also alter physiologic responses to hypocalcemia and possibly exacerbate the problem. 5. Conclusion This retrospective study provides additional evidence that early postoperative PTH levels can accurately predict hypocalcemia in most completion or total thyroidectomies. In this series, the percentage of decrease in PTH levels from preoperatively to 6-hour postoperatively proved to be the most indicative measure. Patients with a greater than 44% decrease in PTH levels were significantly more likely to develop hypocalcemia. We would propose that these patients be provided with supplementation and further inpatient management before discharge. Likewise, patients with a less than 44% decrease in PTH levels during this period are likely safe for discharge from a hypocalcemia standpoint. Further studies will be necessary to prospectively evaluate if the expense of PTH testing can be validated by its use in preventing complications of hypocalcemia and facilitating safe discharge.
References [1] Gonclaves FJ, Kowalski LP. Surgical complications after thyroid surgery performed in a cancer hospital. Otolaryngol Head Neck Surg 2005;132:490-4. [2] Roh JL, Park CI. Routine oral calcium and vitamin D supplements for prevention of hypocalcemia after total thyroidectomy. Am J Surg 2006;192:675-8. [3] Adams J, Andersen P, Everts E, et al. Early postoperative calcium levels as predictors of hypocalcemia. Laryngoscope 1998;108: 1839-41. [4] Nahas ZS, Farrag TY, Lin FR, et al. A safe and cost-effective short hospital stay protocol to identify patients at low risk for the development of significant hypocalcemia after total thyroidectomy. Laryngoscope 2006;116:906-10. [5] Noordzij JP, Lee SL, Bernet VJ, et al. Early prediction of hypocalcemia after thyroidectomy using parathyroid hormone: an analysis of pooled individual patient data from nine observational studies. J Am Coll Surg 2007;205:748-54. [6] Vescan A, Witterick I, Freeman J. Parathyroid hormone as a predictor of hypocalcemia after thyroidectomy. Laryngoscope 2005;115:2105-8. [7] Scurry WC, Beus KS, Hollenbeak CS, et al. Perioperative parathyroid hormone assay for diagnosis and management of postthyroidectomy hypocalcemia. Laryngoscope 2005;115:1362-6. [8] Payne RJ, Hier MP, Tamilia M, et al. Same-day discharge after total thyroidectomy: the value of 6-hour serum parathyroid hormone and calcium levels. Head & Neck 2005;27:1-7. [9] Payne RJ, Tewfik MA, Hier MP, et al. Benefits resulting from 1 and 6hour parathyroid hormone and calcium levels after thyroidectomy. Otolaryngol Head Neck Surg 2005;133:386-90. [10] Snyder SK, Hamid KS, Roberson CR, et al. Outpatient thyroidectomy is safe and reasonable: experience with more than 1,000 planned outpatient procedures. J Am Coll Surg 2010;210:575-84. [11] Sabour S, Manders E, Steward DL. The role of rapid PACU parathyroid hormone in reducing post-thyroidectomy hypocalcemia. Otolaryngol Head Neck Surg 2009;141:727-9. [12] Grodski S, Serpell J. Evidence for the role of perioperative PTH measurement after total thyroidectomy as a predictor of hypocalcemia. World J Surg 2008;32:1367-73. [13] Baird GS, Rainey PM, Wener M, et al. Reducing routine ionized calcium measurement. Clin Chem 2009;55:533-40.
American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216 – 220 www.elsevier.com/locate/amjoto
Parathyroid hormone early percent change: an individualized approach to predict postthyroidectomy hypocalcemia☆ D. Brandon Chapman, MD a , Christopher C. French, MD a , Xiaoyan Leng, MD b , J. Dale Browne, MD a , Joshua D. Waltonen, MD a , Christopher A. Sullivan, MD a,⁎ a
Department of Otolaryngology/Head and Neck Surgery, Wake Forest University School of Medicine, Winston Salem, NC, USA b Department of Biostatistical Sciences, Wake Forest University School of Medicine, Winston Salem, NC, USA Received 21 March 2011
Abstract
Purpose: The purpose of this study is to evaluate a percent change model of postoperative parathyroid hormone level in thyroidectomy patients as a predictor of hypocalcemia. Materials and methods: Chart review was completed on patients who had undergone total or completion thyroidectomy over a 22-month period in our department. Only those patients with a preoperative ionized calcium and parathyroid hormone (PTH) level and at least 1 postoperative result were included. Ionized calcium levels served as an internal control. The Student t test was used to compare PTH level between the normocalcemic and hypocalcemic groups at each time point. Logistic regression analysis was used to predict hypocalcemia based on the diagnostic criteria. Receiver operator curves were used to maximize sensitivity. Results: Fifty-two patients met the inclusion criteria during the study period. A total of 22 patients (42%) experienced hypocalcemia. We were unable to maximize both sensitivity and specificity at the same time point. When comparing preoperative to 6-hour postoperative PTH percent change, patients with a greater than 44% decrease are likely to have hypocalcemia, with a sensitivity of 100%. Likewise, in those patients without a greater than 44% decrease at 6 hours, early discharge can be considered safe, given the negative predictive value of 100%. Conclusion: In our series, patients with a greater than 44% PTH decrease from preoperative to 6hour postoperative are very likely to develop hypocalcemia. We would propose that these patients need further inpatient monitoring to progress to safe discharge. Likewise, patients with a less than 44% decrease at the 6-hour time point are unlikely to develop hypocalcemia and may be considered safe for discharge. © 2012 Elsevier Inc. All rights reserved.
1. Introduction Hypocalcemia is the most common etiology for prolonged hospitalization in patients undergoing total thyroidectomy [1]. Traditional methods for detecting hypocalcemia in the postoperative period include frequent monitoring of total and ionized calcium (ICa) levels as well as close clinical ☆ The authors have no conflicts of interest or external funding sources to disclose. ⁎ Corresponding author. Department of Otolaryngology, 4th Floor Watlington Hall, One Medical Center Blvd, Winston Salem NC, 27157, USA. E-mail address: [email protected] (C.A. Sullivan).
0196-0709/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.amjoto.2011.06.004
monitoring for symptoms of hypocalcemia. These traditional methods can take 24 hours or longer to reveal hypocalcemia, which can prolong hospitalization. A clinical laboratory method for early prediction of postoperative hypocalcemia could, therefore, facilitate earlier discharge (≤24 hours) and improve outcomes after total thyroidectomy. The risks of postoperative hematoma and airway compromise are highest during the first 24 hours after thyroidectomy. However, symptomatic hypocalcemia may not manifest itself during the first 24 hours [2]. With the advent of pay-for-performance measures, many endocrine surgeons are interested in early predication of hypocalcemia, both to prevent serious complications and to begin early treatment in an effort to hasten hospital discharge.
D.B. Chapman et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216–220
Over the past 20 years, clinical research has focused on determining an efficient method for monitoring and predicting postoperative hypocalcemia. Initial efforts focused on early perioperative serum calcium slopes; however, these ultimately proved to have limited negative predictive values (NPVs) [3,4]. In recent years, multiple retrospective studies have emerged, which support the use of postoperative serum parathyroid hormone (PTH) levels as accurate predictors of hypocalcemia in postoperative thyroidectomy patients. A recent meta-analysis examined the pooled data from 9 individual studies examining postoperative PTH levels as predictors of hypocalcemia. This demonstrated that PTH levels at 6 hours postoperatively had a sensitivity of 96.4% and a specificity of 91.4% for detecting postoperative hypocalcemia [5]. Several studies have attempted to correlate PTH levels with clinical outcome while still using calcium levels and clinical picture as the criterion standard. Vescan et al [6] performed a prospective study of 201 patients undergoing total thyroidectomy. Postoperative PTH levels were assessed 1 hour after the procedure and also the following morning. Standard calcium levels and symptom monitoring were used to determine disposition. Results suggested that patients with PTH levels greater than 1.6 pmol/L at 1 hour were likely to remain normocalcemic. Patients with PTH levels less than 1.1 pmol/L at 1 hour were more at risk for developing hypocalcemia. Scurry et al [7] also evaluated postoperative PTH levels in thyroidectomy patients. They determined that accurate predictors for the development of symptomatic hypocalcemia included a decrease in PTH by 75% as well as an absolute postoperative PTH of 7 pg/mL. Payne et al [8,9] have published considerable research on this topic as well. In several articles, they have demonstrated that a PTH greater than 28 ng/L (pg/mL) combined with a normal calcium level at 6 hours postthyroidectomy can be predictive of patients that will maintain normocalcemia. They accomplished a specificity of 100%, sensitivity of 67.9%, and 100% positive predictive value. Based on the aforementioned studies, we have been using patient PTH levels in the perioperative period to assist in decision making and discharge planning since January 2008. However, because of differing reports and variable conclusions in the literature, we have only been able to use these data as an adjunct to clinical decision making while still relying on the criterion standard of ICa levels to predict postoperative hypocalcemia. We have observed that most reports suggest using an absolute cutoff value as a predictor of potential hypocalcemia. We find this approach to be problematic due to the wide range of reporting units (picomolar, picogram per milliliter, and nanogram per liter) and the existence of inherent variability between the assay type/methodology and logistics at different clinical laboratories. Preoperative PTH levels might also be an important consideration, especially in patients with abnormal preoperative parathyroid function. As such, the overarching goal of our study was to examine the
217
use of a percent change model for serum PTH level to predict postoperative hypocalcemia after thyroidectomy.
2. Materials and methods 2.1. Study design After obtaining institutional review board approval, we retrospectively reviewed hospital and clinic records of patients who had undergone total or completion thyroidectomy in our department from January 1, 2008, to October 30, 2009. Primary surgeons included the 3 senior authors (JDB, CAS, and JDW). Ionized calcium levels were used for the internal control group. Intact PTH levels (PTH) were obtained using a standard commercial immunoassay, which measures 1.84 intact PTH molecule (institutional reference range, 12–72 pg/mL). Parathyroid hormone levels were evaluated at the following time points: preoperatively (any time from 30 days before the procedure to immediately before incision), immediately postoperative (exactly 10 minutes after passing off the surgical specimen), and 6 hours postoperatively. Ionized calcium levels were obtained at the following time points: preoperatively; immediately postoperative; and then 6, 12, 24, 36, and 48 hours postoperatively, followed by daily levels for the remainder of hospitalization. These values were rarely obtained after the 24-hour time point, unless the patient remained in the hospital secondary to high drain output, unresponsive or worsening hypocalcemia, or other medical comorbidities. Patients experiencing symptoms of hypocalcemia or with ICa less than 1.00 mmol/L (institutional reference range, 1.00–1.30 mmol/L) were started on oral calcium supplementation and calcitriol immediately. Patients with significant hypocalcemia (ICa b0.80 mmol/L, severe symptoms, or known cardiac conduction problems) were treated with intravenous calcium gluconate, in addition to initiation of oral calcium supplementation and calcitriol. 2.2. Inclusion criteria To be included, the patient required both preoperative and at least 1 set of postoperative PTH and ICa levels. Data reviewed included patient age, sex, date of surgery, primary surgeon, procedure performed, pathologic diagnosis, laboratory values (ICa and PTH), and symptoms. 2.3. Exclusion criteria Patients were excluded from the study if they were undergoing an extended procedure such as total laryngectomy or parathyroidectomy. They were also excluded if they were taking preoperative osteoporotic medications or prophylactic vitamin D and calcium supplementation. Patients with preexisting chronic renal disease, hyperparathyroidism, or known digestive disease were also excluded.
218
D.B. Chapman et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216–220
2.4. Outcome measures Our primary outcome measure was the development of hypocalcemia, as predicted by percent change in PTH level. Hypocalcemia was defined as either a fall in the patient's ICa level below 1.00 mmol/L or signs/symptoms of hypocalcemia during the postoperative period (up to 14 days postoperatively). Commonly accepted signs of hypocalcemia include the Chvostek sign, Trousseau sign, and cardiac arrhythmias. Common symptoms of hypocalcemia that were assessed include perioral paresthesias, distal digital paresthesias, tetany, and palpitations (as documented in the patient's daily progress notes or postoperative clinic notes). 2.5. Statistical analysis Patient demographics and basic characteristics were presented as mean/SD for continuous variables such as age. Patient demographics and basic characteristics were presented as percentages for categorical variables such as sex, procedure, pathologic diagnosis, signs/symptoms of hypocalcemia, and hypocalcemia proven by laboratory values. Mean PTH levels were calculated at each time point (preoperatively, immediately postoperatively, and 6hour postoperatively), and the 2-sided Student t test was performed to compare the normocalcemic and hypocalcemic groups. Logistic regression was used to predict hypocalcemia based on the diagnostic criteria. Sensitivity and specificity were then calculated based on different cutoff points from the logistic regression model. Receiver operator curves were plotted with sensitivity vs 1-specificity. The cutoff point with maximum sensitivity was chosen as optimal for this screening test. Diagnosis criteria included both absolute PTH change and percent change for the immediate postoperative time point and the 6-hours postoperative time point. Sex effect was also examined using logistic regression model.
3. Results There were a total of 320 thyroidectomy cases performed by the Wake Forest University Otolaryngology service, which were reviewed for inclusion criteria over a period of 20 months. A total of 52 (16% of all reviewed) patients met the inclusion criteria during the study period. The most common reasons for exclusion included a hemithyroidectomy, preoperative calcium supplementation, and lack of required laboratory values at the specified time points. Mean age of our study population was 50.2 years, and 69% were female. Papillary carcinoma was identified in 58% of patients. Most patients (94%) underwent total thyroidectomy (vs completion thyroidectomy). Central compartment neck dissection was included in the procedure in 13 (25%) of the patients. A total of 22 patients (42%) experienced hypocalcemia (based on symptoms or ICa level). The complete data set including patient age, sex, procedure performed, final
pathologic diagnosis, incidence of signs/symptoms of hypocalcemia, central neck compartment dissections, and incidence of hypocalcemia is displayed in Table 1. The mean preoperative PTH was 69.0 pg/mL, with no significant difference between the normocalcemic and hypocalcemic groups (P = .2317). However, a significant difference existed in mean PTH levels when comparing the normocalcemic and hypocalcemic groups during the immediate postoperative and 6-hour postoperative time points (P = .0144 and P = .0087, respectively). Mean values for each group are included in Table 2. Sex was not found to have effect on the incidence of hypocalcemia using either the absolute change or percent change models. We were unable to maximize both sensitivity and specificity at the same time point. Therefore, given that the main goal of this study was to establish an accurate screening test, we calculated a cutoff point that could independently maximize the sensitivity. Receiver operator curves were used to achieve this task. It was found that within the immediate postoperative period, results were too variable to be applicable to the study goals. The 6-hour time point appeared optimal for maximizing sensitivity. This analysis (Table 3) revealed that when comparing the percentage of change from preoperative to 6-hour postoperative PTH levels, patients with a greater than 44% decrease were more likely to have hypocalcemia (sensitivity, 100%) (Fig. 1). Likewise, patients having a
Table 1 Patient demographics and basic characteristics Study group
N = 52
Mean age in years (SD) Sex Male Female Procedure Total thyroidectomy Completion thyroidectomy Pathologic diagnosis Papillary carcinoma Multinodular goiter Hashimoto's thyroiditis Medullary carcinoma Follicular adenoma Hurthle cell adenoma Thyroglossal duct cyst Signs/symptoms of hypocalcemia Digital paresthesia Chvostek sign Palpitations Perioral paresthesia None Central neck compartment dissections Yes No Hypocalcemia (ICa b1.0, or Si/Sx) Yes No
50.2 (18.0) 16 (31%) 36 (69%) 49 (94.2%) 3 (5.8%) 30 (58%) 12 (23%) 5 (10%) 2 (4%) 1 (2%) 1 (2%) 1 (2%) 4 (8%) 1 (2%) 1 (2%) 0 (0%) 46 (88%) 13 (25%) 39 (75%) 22 (42%) 30 (58%)
D.B. Chapman et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216–220
219
Table 2 Mean differences in PTH at different time points
Preoperative Immediate postoperative 6-h postoperative
All patients
Normocalcemic
Hypocalcemic
P⁎
69.0 ± 39.9 27.7 ± 23.8
63.3 ± 41.5 34.7 ± 23.8
76.8 ± 37.2 18.5 ± 21.0
.2317 .0144
24.3 ± 23.3
32.2 ± 20.7
14.9 ± 23.1
.0087
Data presented as mean ± SD (picogram per milliliter). ⁎ P value compares normocalcemic mean PTH to hypocalcemic mean PTH.
less than 44% decrease at 6 hours could be considered safe for early discharge, given the NPV of 100%.
4. Discussion Traditional serum calcium monitoring is highly specific, yet this is also time consuming. Because of the physiologic time lag between laboratory draws and the manifestations of hypocalcemia, this method is intrinsically flawed and will always demonstrate low sensitivity during early serum calcium measurements. It was our goal to determine if early percent change in serum PTH levels could prove to be an early and accurate predictor of hypocalcemia. The ideal screening test maximizes sensitivity so that no patients are “missed.” Using the percent change model to accurately predict eventual hypocalcemia, this study demonstrated a maximum sensitivity when the change in PTH levels from preoperatively to 6-hour postoperatively decreased by 44% or greater. This can aid in early identification of patients who will likely develop hypocalcemia and, thus, will likely benefit from early supplementation. After beginning supplementation, we recommend traditional calcium testing because PTH levels would be skewed by exogenous calcium. Likewise, this study demonstrated an NPV of 100% when comparing the preoperative and 6-hour postoperative levels. This infers that patients not exceeding a 44% drop in PTH level at 6 hours will likely not develop hypocalcemia. We would propose that these patients can be safely discharged once the risk for other perioperative complications has passed. These patients should be given warnings for the
Table 3 Diagnosis criteria comparisons, maximizing sensitivity Test
Pre—6-h absolute change (pg/mL)
Pre–6-h postoperative % change
Cutoff value Sensitivity Specificity Positive predictive value NPV
−24 1 0.423 0.595
−44 1 0.538 0.647
1
1
Fig. 1. Six-hour percent change vs sensitivity.
signs and symptoms of hypocalcemia but would likely not need additional laboratory draws. Our optimization of sensitivity comes at the expense of specificity. Although the percent change model can be established as an excellent screening test, one must keep in mind that there will be a higher rate of false positives. Although inappropriate treatment with calcium and vitamin D analogs can lead to hypercalcemia and subsequent symptom, short-term “overtreatment” is, in the authors' minds, less risky than undertreatment. We would prefer to capture all potentially hypocalcemic patients, as we have seen more severe complications from undertreated patients than from overtreated patients. Many institutions [10], including our own, have in the past routinely treated patients with prophylactic oral calcium and vitamin D with a low incidence of complications from hypercalcemia. With that in mind, careful postoperative follow-up and rapid weaning of supplementation should be sought. Although there has been an abundance of retrospective studies examining the predictive powers of postoperative PTH levels, true outcome studies have been limited. Sabour et al [11] retrospectively demonstrated that the use of rapid post anesthesia care unit (PACU) PTH levels in decision making for supplementation can impact the overall rate of hypocalcemia in thyroidectomy patients. They discovered that use of immediate PTH levels in the PACU significantly reduced the proportion of patients experiencing both mild and severe hypocalcemia but did not completely prevent hypocalcemia in all patients. In this cohort, decisions were based on absolute values of PTH. A recent literature review by Grodski and Serpell [12] reviewed 27 series and concluded that a single PTH measurement drawn any time from 10 minutes postoperatively to several hours later will provide equally accurate results. This study is not without limitations. The retrospective nature of this study makes it difficult to ensure complete capture of the signs and symptoms of hypocalcemia. The low rate of patient capture due to missing laboratory values is also a reflection of the retrospective nature. In addition, there is little consensus in the literature as to what ICa levels should prompt intervention or what intervention should be undertaken [13]. Our definition of hypocalcemia, based on ICa less than 1.00
220
D.B. Chapman et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 216–220
mmol/L, may capture patients who would not have developed signs or symptoms of hypocalcemia. Our preference, however, is to favor the capturing of false positives, rather than miss potential hypocalcemic patients. The retrospective nature of this study also limited our ability to ensure preoperative PTH and ICa laboratory evaluation, which significantly reduced our sample size due to exclusion criteria. Another limitation that has plagued many similar studies has been a difficulty in the classification of patients with regard to their preoperative risk factors, specifically related to their medication use. Many patients are on over-thecounter regimens including calcium and vitamin D, which is often difficult to capture in a retrospective series. This is further confounded by the increasing use of bisphosphonates for osteoporosis, which can also alter physiologic responses to hypocalcemia and possibly exacerbate the problem. 5. Conclusion This retrospective study provides additional evidence that early postoperative PTH levels can accurately predict hypocalcemia in most completion or total thyroidectomies. In this series, the percentage of decrease in PTH levels from preoperatively to 6-hour postoperatively proved to be the most indicative measure. Patients with a greater than 44% decrease in PTH levels were significantly more likely to develop hypocalcemia. We would propose that these patients be provided with supplementation and further inpatient management before discharge. Likewise, patients with a less than 44% decrease in PTH levels during this period are likely safe for discharge from a hypocalcemia standpoint. Further studies will be necessary to prospectively evaluate if the expense of PTH testing can be validated by its use in preventing complications of hypocalcemia and facilitating safe discharge.
References [1] Gonclaves FJ, Kowalski LP. Surgical complications after thyroid surgery performed in a cancer hospital. Otolaryngol Head Neck Surg 2005;132:490-4. [2] Roh JL, Park CI. Routine oral calcium and vitamin D supplements for prevention of hypocalcemia after total thyroidectomy. Am J Surg 2006;192:675-8. [3] Adams J, Andersen P, Everts E, et al. Early postoperative calcium levels as predictors of hypocalcemia. Laryngoscope 1998;108: 1839-41. [4] Nahas ZS, Farrag TY, Lin FR, et al. A safe and cost-effective short hospital stay protocol to identify patients at low risk for the development of significant hypocalcemia after total thyroidectomy. Laryngoscope 2006;116:906-10. [5] Noordzij JP, Lee SL, Bernet VJ, et al. Early prediction of hypocalcemia after thyroidectomy using parathyroid hormone: an analysis of pooled individual patient data from nine observational studies. J Am Coll Surg 2007;205:748-54. [6] Vescan A, Witterick I, Freeman J. Parathyroid hormone as a predictor of hypocalcemia after thyroidectomy. Laryngoscope 2005;115:2105-8. [7] Scurry WC, Beus KS, Hollenbeak CS, et al. Perioperative parathyroid hormone assay for diagnosis and management of postthyroidectomy hypocalcemia. Laryngoscope 2005;115:1362-6. [8] Payne RJ, Hier MP, Tamilia M, et al. Same-day discharge after total thyroidectomy: the value of 6-hour serum parathyroid hormone and calcium levels. Head & Neck 2005;27:1-7. [9] Payne RJ, Tewfik MA, Hier MP, et al. Benefits resulting from 1 and 6hour parathyroid hormone and calcium levels after thyroidectomy. Otolaryngol Head Neck Surg 2005;133:386-90. [10] Snyder SK, Hamid KS, Roberson CR, et al. Outpatient thyroidectomy is safe and reasonable: experience with more than 1,000 planned outpatient procedures. J Am Coll Surg 2010;210:575-84. [11] Sabour S, Manders E, Steward DL. The role of rapid PACU parathyroid hormone in reducing post-thyroidectomy hypocalcemia. Otolaryngol Head Neck Surg 2009;141:727-9. [12] Grodski S, Serpell J. Evidence for the role of perioperative PTH measurement after total thyroidectomy as a predictor of hypocalcemia. World J Surg 2008;32:1367-73. [13] Baird GS, Rainey PM, Wener M, et al. Reducing routine ionized calcium measurement. Clin Chem 2009;55:533-40.