Diagnostic Accuracy Study of Intraoperative and Perioperative Serum Intact PTH Level for Successful Parathyroidectomy in 501 Secondary Hyperparathyroidism Patients

Parathyroidectomy (PTX) is an effective treatment for severe secondary hyperparathyroidism (SHPT); however, persistent SHPT may occur because of supernumerary and ectopic parathyroids. Here a diagnostic accuracy study of intraoperative and perioperative serum intact parathyroid hormone (iPTH) was performed to predict successful surgery in 501 patients, who received total PTX + autotransplantation without thymectomy. Serum iPTH values before incision (io-iPTH0), 10 and 20 min after removing the last parathyroid (io-iPTH10, io-iPTH20), and the first and fourth day after PTX (D1-iPTH, D4-iPTH) were recoded. Patients whose serum iPTH was >50 pg/mL at the first postoperative week were followed up within six months. Successful PTX was defined if iPTH was <300 pg/mL, on the contrary, persistent SHPT was regarded. There were 86.4% patients underwent successful PTX, 9.8% remained as persistent SHPT and 3.8% were undetermined. Intraoperative serum iPTH demonstrated no significant differences in two subgroups with or without chronic hepatitis. Receiver operating characteristic (ROC) curves showed that >88.9% of io-iPTH20% could predict successful PTX (area under the curve [AUC] 0.909, sensitivity 78.6%, specificity 88.5%), thereby avoiding unnecessary exploration to reduce operative complications. D4-iPTH >147.4 pg/mL could predict persistent SHPT (AUC 0.998, sensitivity 100%, specificity 99.5%), so that medical intervention or reoperation start timely.


Number of resected parathyroid glands in PTX patients. Operative exploration revealed that most
patients had four parathyroid glands (435, 86.8%). Five glands were resected in 2 patients (0.4%) who turned out to have successful PTX. Three glands were identified in 49 patients (9.8%) and two glands were removed in 15 patients (3.0%) after careful exploration. In fact, the results in persistent SHPT group did not actually reflect their parathyroid gland numbers. Persistent SHPT patients tended to have more parathyroid glands than the recorded numbers (Fig. 2).
Intraoperative and perioperative iPTH levels in PTX patients. Overall, the mean percentage reduction of serum iPTH at 10 and 20 minutes after PTX (io-iPTH10%, io-iPTH20%) in the successful PTX group were 86.8% and 90.6%, respectively, while the persistent SHPT group were 69.8% and 76.5%. We compared io-iPTH10% and io-iPTH20% between the successful PTX group and persistent SHPT group and found significant differences (both P< 0.001).
There were 204 hepatitis patients and 297 non-hepatitis patients in the PTX group. The baseline iPTH levels of patients with or without chronic hepatitis were 2,009.9 (1,393.5-2,801.4) pg/mL and 2,163.9 (1,577.9-3,009.9) pg/mL, respectively (P = 0.116). The levels of serum iPTH at 10 or 20 minutes after PTX (io-iPTH10, io-iPTH20) and their percentage reductions in these two subgroups also showed no significant differences (Fig. 3).
Intraoperative and perioperative iPTH cutoff values for the prediction of successful PTX or persistent SHPT. The cutoff values of intraoperative and perioperative iPTH were listed in Table 2. Sensitivity and specificity were obtained from the ROC analyses (Fig. 5). It showed that the criterion of an 82.9% decline in io-iPTH10% could predict the success of surgery (area under the curve [AUC] 0.857, sensitivity 85.5%, and specificity 73.1%). In addition, 88.9% decline of io-iPTH20% was effective to predict successful PTX (AUC 0.909, sensitivity 78.6%, and specitivity 88.5%). D1-iPTH could predict persistent SHPT with a cutoff value of 100.5 pg/mL. AUC, sensitivity, and specitivity of this criterion were 0.999, 100% and 98.6% respectively. When D4-iPTH was used to predict persistent SHPT, the cut-off value was 147.4 pg/mL, the AUC was 0.998, sensitivity was 100%, and specificity was 99.5%.

Discussion
Successful PTX could improve symptoms effectively and reduce the risk of all-cause and cardiovascular mortality in severe SHPT patients 14 . All parathyroid glands must be removed completely during the surgery for successful PTX. SHPT patients had a higher risk of supernumerary parathyroid glands 5,15 . Our results demonstrated that the rate of resected 2, 3, 4, or 5 parathyroids were 3.0%, 9.8%, 86.8% and 0.4%, respectively. Actually, the exact number of parathyroids in persistent SHPT patients was more than we recorded.    Parathyroid glands originate from the endoderm of the third and fourth pharyngeal pouches. Based on the anatomic and developmental characteristics of parathyroid glands, surgical procedures for SHPT patients include subtotal parathyroidectomy (sPTX) and total parathyroidectomy (tPTX) with or without autotransplantation (AT). Ectopic parathyroid could exist in any location of the migration path including the intrathyroid, carotid sheath, thymus, and upper mediastinum 13 . According to previous data, the frequency of ectopic parathyroid glands was about 15% in SHPT patients [16][17][18] . Supernumerary and ectopic parathyroid glands make it difficult to perform successful PTX. Furthermore, excess exploration carries a higher risk of surgical complications including nerve injuries and bleeding, also increase perioperative period mortality. Thus, the confirmation of the complete removal of parathyroid glands during the surgery is required. Our results demonstrated that io-iPTH monitoring was a useful tool for predicting successful PTX in SHPT patients (Fig. 5). We showed more than an 82.9% decrease of io-iPTH10 could predict a complete parathyroid gland excision. As Fig. 4 shows, iPTH levels decreased gradually during the surgery. If io-iPTH10 failed to reach the above criterion, we recommend a criterion of > 88.9% decrease in io-iPTH20. Also, we studied the perioperative iPTH values and found that D4-iPTH exceeding 147.4 pg/mL could effectively predict the persistence of SHPT after surgery. Patients not achieving this criterion are suggested to be followed-up closely, and medication intervention should be started. Another operation may even be necessary.  Studies identified the liver as a major extra renal site of iPTH metabolism 6,19 . Chronic hepatitis, which is mainly caused by hepatitis B virus (HBV) and hepatitis C virus (HCV), could be transmitted via infected blood products. Among patients receiving maintenance dialysis, the prevalence of HBV ranges from 1.3-14.6% and the prevalence of HCV is 0.7-18.1% 20,21 . To our knowledge, the influence of chronic hepatitis on intraoperative serum iPTH values in PTX patients has not been studied previously. We showed that hepatitis had no effect on intraoperative iPTH values. Perhaps this may be explained by extremely abnormal hepatic functions being a contraindication of surgery. Our patients all had normal serum ALT, AST, DBIL, IBIL levels and prothrombin times even if they had a chronic hepatitis history.  Hiramitsu et al. 22 studied the predictive value of io-iPTH in 226 PTX patients and found that an iPTH value of < 60 pg/mL on postoperative day 1 (sensitivity and specificity not mentioned) could predict successful parathyroidectomy and a 70% io-iPTH drop from the baseline at 10 minutes after surgery (sensitivity 97.5%, specificity 52.2%) could determine sufficient parathyroid gland removal. Compared with Hiramitsu's research, the sensitivity and specificity of our criterion were higher, although there were some similarities. Our sample size was larger than theirs (501 vs 226). Further, we reported novel findings that there was no influence of chronic hepatitis on intraoperative iPTH values. In addition, we suggested that D4-iPTH > 147. 4    99.5%) as a criterion of persistent SHPT could help physicians made an accurate decision to start medical treatments in a timely manner.
In conclusion, our results showed that chronic hepatitis had no effect on intraoperative iPTH monitoring. More than 88.9% of io-iPTH20% was effective to predict the success of surgery (sensitivity 78.6%, specificity 88.5%) and D4-iPTH > 147.4 pg/mL was effective to predict persistent SHPT (sensitivity 100%, specificity 99.5%). Intraoperative iPTH monitoring could help surgeons make decisions to stop exploration in a timely manner, which will obviously reduce the risk of bleeding and nerve injuries obviously. Perioperative iPTH monitoring could effectively predict the persistence of SHPT after surgery and the patients can start medical intervention or reoperation as soon as possible.
Concise Methods PTX patients. We enrolled 501 PTX patients including 277 men and 224 women who received total parathyroidectomy with forearm autotransplantation (tPTX + AT) without thymectomy in our hospital from April 2011 to August 2015. All data were collected retrospectively. PTX was performed in severe SHPT patients who failed to respond to medical therapy 2 . Our surgical indications included: persistent serum iPTH > 800 pg/mL; hypercalcemia and/or hyperphosphatemia that could not be controlled by medical therapy; obvious clinical manifestations such as bone pain, pruritus, ectopic calcification or fracture; and at least one enlarged parathyroid gland discovered by ultrasound or a radiopharmaceutical technetium-99m-methoxyisobutylisonitrile (99mTc-MIBI) scan. None of the patients took vitamin D analogs and calcimimetics.
Healthy and stage 5 CKD group. We also included 120 healthy volunteers and 121 stage 5 CKD patients without severe SHPT.
All clinical investigations were conducted according to the 2008 Declaration of Helsinki and good clinical practice guidelines. Written informed consent was obtained from all the subjects. The study protocols were approved by the Research Ethics Committee of the First Affiliated Hospital of Nanjing Medical University, People's Republic of China.
Surgical procedure. Preoperative evaluations included neck ultrasonography and parathyroid scintigraphy (99mTc-MIBI) for demonstrating the number, size, and location of parathyroid glands. Pulmonary functions, cardiac function, routine blood tests, and coagulation tests (including prothrombin time) were conventionally conducted before surgery so that contraindications would be discovered. TPTX + AT without thymetomy was performed routinely under general anesthesia in all SHPT patients. All operations were performed by the same surgeon. Bilateral neck examinations were carefully performed to make sure all hyperplastic parathyroid glands were resected.
Intraoperative frozen section analysis was routinely adopted to verify that the resected specimen was parathyroid tissue. The selected diffuse hyperplasia parathyroid fragment was cut into slices about 1 mm 3 and 8 slices were transplanted into forearm muscles without an arteriovenous fistula for hemodialysis. After surgery, pathological sections were examined carefully.

Definition of successful PTX and persistent SHPT.
According to previous studies, serum iPTH levels detected at the first postoperative week < 300 pg/mL was the criterion of successful PTX 13 . Here we adopted a stricter criterion. Patients with serum iPTH< 50 pg/mL at the first postoperative week were classified as successful PTX group. Patients with serum iPTH > 50 pg/mL at the first postoperative week examination were followed up to verify the effect of surgery. Depending on serum iPTH values within six months, patients with iPTH < 300 pg/mL were regarded as the successful PTX group, and those whose iPTH were > 300 pg/mL were classified as persistent SHPT. Patients without valid iPTH values after surgery were classified as the undetermined group.
Statistical analysis. All statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 20.0 (SPSS Inc., Chicago, IL, U.S.). Continuous variables were presented as mean ± SD or median (interquartile range), and categorical variables were presented as number and proportion. Differences between groups were compared using an independent samples t or Wilcoxon rank sum test for continuous variables and a chi-squared or Fisher's exact test for categorical variables. P < 0.05 was considered statistically significant. Receiver operating characteristic (ROC) curves were used to identify the cutoff value for prediction of surgical success and persistent SHPT. Diagnostic accuracy was expressed through sensitivity, specificity, and the area under the ROC curve (AUC). The Standard Reporting for Diagnostic (STARD) studies were used here.
Patients were designated as true positive (TP), true negative (TN), false positive (FP), and false negative (FN) ( Table 3). The PPV and NPV were calculated.