Hemithyroidectomy, does the indication influence the outcome?

Nodules in the thyroid gland are common [1]. Patients may discover the nodules themselves or experience compression symptoms by them such as neck fullness, dysphagia, odynophagia, coughing, dyspnea, and voice changes as a result of the anatomy of the neck [2, 3]. A thyroid nodule may also secrete extensive levels of thyroid hormones as a solitary toxic adenoma causing hyperthyroidism [4]. Preoperative investigations include ultrasound and fine-needle aspiration where the cell formations are evaluated and classified according to scoring systems (THY, Bethesda) [5, 6]. The higher the score, the greater the risk of malignancy. Scintigraphy may also be indicated in order to determine uninodular/multinodular disease. Whether surgery is necessary or not is determined by the patients’ symptoms, the Bethesda classification, and the ultrasound examination. Patients with a thyroid nodule suspicious for malignancy are recommended hemithyroidectomy for diagnostic purposes in order to confirm or rule out a cancer diagnosis [6]. Patients in whom thyroid malignancy is not suspected but who experience compression symptoms in the throat from a predominantly unilateral thyroid enlargement are usually recommended surgery to remove half of the thyroid gland (hemithyroidectomy). Hemithyroidectomy for toxic adenomas is less common. Thus, hemithyroidectomy may be recommended for mainly two different indications, either for investigating a cancer suspicion (“excluding malignancy”) or for relieving compression symptoms.

As with any surgery, thyroidectomy carries the risk of various complications. Following hemithyroidectomy, potential major complications include infection [7, 8], bleeding [8‐10], and injury to the recurrent laryngeal nerve (RLN) [8, 11, 12]. Preoperatively, the patient must therefore be informed regarding the extent of the procedure, risks, and the rationale for the surgery. Awareness of the occurrence rate and potential risk factors is therefore essential for the surgeon, both for evaluation of surgical and medical options. Studies have shown that the extent of resection (subtotal versus total) is an independent risk factor for paresis of the RLN [13, 14]. To minimize vocal cord paralysis, the surgeon’s goal is to visualize the recurrent nerve. Intraoperative nerve monitoring (IONM) has been introduced to help further detection of the RLN and reduce the risk of complication [11, 12]. With this in view, one could consider a hemithyroidectomy to be rather extensive when it is performed to exclude malignancy, since the majority (70–95%) of these patients have no symptoms and will have benign disease [15]. However, due to the lack of other definitive diagnostic options, this approach is widely used and accepted.

The risk of complications has not been studied in detail in patients undergoing diagnostic hemithyroidectomy. It could be possible that the risk for paresis RLN is particularly high in this group because surgeons might have a tendency to be as radical as possible when operating on patients with suspected cancer. We sought to investigate this hypothesis. As a control group, we chose a patient group which is also recommended hemithyroidectomy though due to compression symptoms.

Our results show that hemithyroidectomies to exclude malignancy overall are associated with a very low risk for complications for the patients.

The strength of this study was a large study sample size including 9677 patients that were operated solely with hemithyroidectomy from a Scandinavian quality register. Nearly, all the variables have been the same since the start of the register in 2004, except for IONM where the registration began in 2009. The register contains many variables with several follow-ups and is validated annually. There is a continual discussion regarding routing for input and follow-up. It should be taken into consideration that despite a small number of complication rates, we found statistically significant differences because of the large sample size, so these findings should be considered as trends rather than significance.

Unfortunately, there is no classification system such as Bethesda or THY recorded on the register. If so, then the patients could have been stratified into categories, evaluating if it was an accurate indication to perform hemithyroidectomy. If the surgeon makes an overall assessment for an unknown thyroid tumor, then the patient is selected to undergo operation for excluding malignancy. However, the histopathological examination in excluding malignancy group showed a malignancy rate of 14.7% which is equivalent to a Bethesda 3 (5–15%) or Bethesda 4 (15–30%) classification [35] being an adequate indication to diagnostic hemithyroidectomy.

Another challenge with register data is that the researcher often is not involved in the data collection. Errors may occur when entering the data. For instance, we identified a few patients where the indication for surgery was classified as “excluding malignancy” despite the fact that malignancy already had been confirmed prior surgery by fine-needle aspiration cytology. These cases were excluded from the final study population.

Surprisingly, among those patients that were diagnosed with a RLN paresis after the operation, no information was available in 38% of the nerve status at the 6–12-month follow-up. A study using the EUROCRINE register showed that 37 of 50 (74%) patients with a vocal cord paresis diagnosed earlier during follow-up did not undergo a long-term laryngoscopy [36]. Adjustments for this variable were made in our study (Table 3) and showed that the compression symptoms group had more permanent RLN paresis (6–12-month follow-up) than the excluding malignancy group.

Another methodical consideration is that our study population did not undergo systematic laryngoscopy at all the various follow-up times. Therefore, some patients who were unexamined at earlier follow-up times were later confirmed having a paresis. Studies have shown that IONM has a 98–100% [37‐41] negative predictive value when comparing IONM with postoperative laryngoscopy. Therefore, the number of normal RLN in Table 3 was calculated as the number of patients with intact IONM during surgery and no clinical signs/evidence of RLN injury directly after, 6 weeks after, and 6–12 months after surgery.

In conclusion, there is a range of factors that interfere and contribute to bleeding, infections, and RLN paresis after hemithyroidectomy. The frequency of these complications was low, but it was noted that patients operated due to compression symptoms were more likely to receive permanent RLN paresis than those operated to exclude malignancy. In this observational study based on a Scandinavian quality register, the indication “excluding malignancy” for hemithyroidectomy was associated with lower risk of permanent RLN paresis. Patients can however be reassured that hemithyroidectomy is a safe surgical procedure and does not entail an extensive risk.