IP is a common finding in pathology reports of thyroid surgeries. Previous series have described IP rates between 5.8% and 29% (14, 15). The wide range of IP rates may be explained by the different types of thyroid surgery described in those series (total thyroidectomy with or without CLND, subtotal thyroidectomy, lobectomy, isthmusectomy). In studies that focused on total thyroidectomy, the IP rate varied between 16.2% and 22.4% (16, 17). Our study found an avoidable IP rate (excluding intrathyroidal IP) of 15.2%, which are consistent with recent literature reports.
The incidence of intrathyroidal parathyroid glands in cadaver studies is always very low (18, 19). However, the reported incidence of intrathyroidal parathyroid glands on pathology reports of previous studies describing IP varied between 0.2% and 10.2% of all thyroid specimens(11, 12), representing between 2.2% and 57.5% of resected parathyroid glands. We chose to exclude intrathyroidal parathyroid glands to keep the study focus on avoidable IP.
In our study, the incidence of transient hypocalcemia was significantly higher in patients with IP. The association between IP and transient hypocalcemia remains debated, because many studies did not find any statistical link between IP and transient hypocalcemia. However, the primary objective of those studies was to provide a description of IP, and they included patients undergoing less than total thyroidectomy or a small number of total patients (20–22). Recent reports that focused on patients undergoing total thyroidectomy or total thyroidectomy with central neck dissection, mainly found a link between IP and transient hypocalcemia (17, 23, 24).
In our study, patients who had IP were more likely to present with permanent hypoparathyroidism, even those with only 1 incidentally removed parathyroid gland. A 3-fold increased risk of permanent hypoparathyroidism was found in patients with IP after total thyroidectomy. The influence of IP on permanent hypoparathyroidism remains controversial. An association between permanent hypoparathyroidism and IP has been reported in only a few studies (11, 13, 15, 23–25). This may be explained by the low incidence of permanent hypoparathyroidism, or the moderate incidence of IP associated with the insufficient sample size of many studies. Studies reporting an association between permanent hypoparathyroidism and IP had a rather large population (24), an elevated IP rate (15, 23, 25) or an elevated permanent hypoparathyroidism rate (11, 13, 23).
Many studies have attempted to identify the potential risk factors for IP.
Demographic data such as young age and female sex have been reported as risk factors for IP (12, 26, 27). However, this was concluded on univariate analysis and was not verified using multivariate models. In our study, neither age nor sex was different between patients with and without IP.
With respect to the final pathology report, many studies have reported malignancy as a risk factor for IP (17, 24). In our study, the incidence of malignancy was significantly higher in patients with IP on univariate analysis. However, on multivariate analysis, malignancy was not an independent risk factor for IP. We considered that it may be a confounding factor because a large majority of our patients with malignancy had CLND.
Some studies have reported Grave’s disease, chronic thyroiditis, and thyroid weight as risk factors for IP (11, 28, 29, 30). In contrast, in our study, the incidence of Grave’s disease and thyroid weight was significantly lower in patients with IP on univariate analysis, although they were not independent protective factors in multivariate analysis. For Grave’s disease (small thyroid gland in most cases), this result was consistent with at least 1 other published study with a large sample size (11).We hypothesized that as we expected a more difficult dissection and a higher rate of postoperative hypocalcemia in those patients, we actually were more careful about parathyroid gland preservation. On the opposite, we found that IP was more frequent in case of Hashimoto thyroiditis, but not in the multivariate model. We assumed that thyroiditis was involved in postoperative hypocalcemia, but not by increasing the IP rate.
Many previous studies have reported CLND as a risk factor for IP. In our study, CLND was the strongest independent risk factor for IP, with a ≥ 4-fold increased risk of IP in patients with CLND. Therefore, surgeons should be aware of the risk-benefit ratio of CLND for each patient and patients should be adequately informed. We found that if some unintentionally resected thymus tissues were recorded in the central neck dissection in the pathology report, the risk of IP significantly increased. To our knowledge, this is the first study to report the incidental resection of thymus tissue as an independent risk factor for IP. Moreover, the presence of thymus tissue in CLND specimens did not improve the quality of the CLND because the number of resected and positive lymph nodes was the same in the 2 groups. We suggest avoiding resecting any thymus tissue during central neck dissection. Some authors have developed a surgical technique to preserve the thymus during central neck dissection with a significantly improved rate of inferior parathyroid preservation (31).
Strength and limitations
The strengths of our study were that despite its retrospective nature, patient data were recorded consecutively and prospectively with great thoroughness. Calcium and parathormone blood levels were systematically collected and assessed to ensure that all cases of transient and permanent hypocalcemia were recorded. The long-term follow-up allowed us to determine the long-term consequences of IP.
However, the present study had some limitations. Most surgeons did not document the number of identified parathyroid glands or the macroscopic vascularization, precluding any correlation analysis between parathyroid identification and IP. We believe that parathyroid glands do not need to be routinely identified and assessed if they are not encountered during thyroidectomy (33).
We excluded patients whose pathology reports mentioned only “parathyroid tissue” to include only those who had 1 or several whole gland(s) incidentally removed. Because of the variability of pathologists who evaluated the specimens and the quasi-constant lack of size measurements of the resected “parathyroid tissue”, we could have potentially excluded a small number of patients with IP.