Incidence, Risk Factors, and Natural Outcome of Vocal Fold Paresis in 920 Thyroid Operations with Routine Pre- and Postoperative Laryngoscopic Evaluation

Vocal fold paresis (VFP) caused by recurrent laryngeal nerve (RLN) injury is a well-known complication of thyroid surgery, and it has been widely documented in the literature. However, a systematic review by Jeannon and colleagues demonstrated a high variation in screening methods and VFP rates between previously published studies. The rates of transient VFP ranged from 1.4 to 38.4% (mean 9.8%) and from zero to 18.6% (mean 2.3%) for permanent VFP [1]. The incidence of VFP may be underestimated unless routine vocal fold evaluation is performed. In a study involving 26 Scandinavian hospitals with 3660 registered thyroid operations, those institutions that practiced routine postoperative laryngoscopy reported almost two times higher rates of VFP than those that did not [2].

Postoperative RLN injury is considered permanent if complete vocal fold immobility or dysfunction lasts more than 1 year [3]. Permanent injuries have been documented in up to 1.4% and transient injuries in 5.2–12.6% of the patients according to studies which utilized routine postoperative vocal fold examination [4, 5]. The reported risk factors for intraoperative RLN injury include patient’s older age, intrathoracic goiter, thyrotoxicosis, thyroid malignancy, previous thyroidectomy, reoperation for bleeding, extended surgery, low or medium volume hospital, and low-volume surgeons [2, 5‐8].

The primary aim of this study was to prospectively assess the prevalence of incidental preoperative VFP and the incidence rate of perioperative RLN injury using routine laryngoscopy screening before and after thyroid surgery. The secondary aims were to identify risk factors for VFP and to analyze the outcome of postoperative VFP during 12-month follow-up.

During the 6-year study period, 920 thyroid operations were performed in 866 patients (mean age 55 ± 16 years, 82% female) with 1296 nerves at risk (Table 1). Preoperative and postoperative laryngoscopy was performed in 95% and 98% of the cases, respectively. Preoperatively, 24 patients (2.8%) had VFP prior to the index operation; 14 were symptomatic. Six had undergone previous ipsilateral thyroid operation, which was the probable cause of the injury. Eight patients with preoperative VFP had malignant thyroid nodule on the same side as the paresis, whereas ten patients had no known cause for the incidental preoperative VFP and thus were considered idiopathic. Nineteen of the preoperative VFPs persisted, and five resolved during the follow-up after surgery.

Postoperatively, new unilateral VFP was detected after 51 operations. Two patients suffered from new bilateral VFP after surgery; one operation was performed due to recurrent goiter and the indication was hyperthyroidism in the other. Two out of the 51 patients with new unilateral VFP had contralateral VFP preoperatively and therefore bilateral VFP postoperatively. The rates of new VFP were 5.8% (n = 53/920) per operations and 4.2% (n = 55/1296) per nerves at risk.

The indications for surgery with corresponding rates of VFP and malignant pathology findings are presented in Table 2. The rates of new VFP and definitive permanent VFP were 5.2% and 2.9% for primary goiter, 22.5% and 15.0% for recurrent goiter, 3.7% and 1.8% for operations performed due to suspicious thyroid nodule, 20.5% and 12.8% for operations with FNA biopsy verified malignancy, and 4.1% and 0.8% for thyroidectomies due to hyperthyroidism, respectively. Malignant histology was found in 172 (19%) of all 920 surgical pathology specimens. Unexpected malignancies were found in 39/383 (10%) operations performed for symptomatic goiter. Malignant neoplasms were confirmed in 73/271 (27%) cases operated for suspicious or undetermined thyroid nodule. Moreover, 17 additional malignancies were found in 56 completion thyroidectomies (30% incidence rate for completion procedures). When FNA biopsy revealed a high suspicion of malignancy, the postoperative histologic examination confirmed carcinoma in 97% of cases (one proved to be a follicular adenoma).

Although well known and widely reported in the literature, the incidence of VFP before and after thyroid surgery is exceptionally varying. An important feature of the present study was that the decision to screen for VFP did not rely on the patient’s or surgeon’s judgment. Instead, nearly all cases were examined independently by third-party investigators (otolaryngologists) with 95% preoperative and 98% postoperative laryngoscopic screening coverage.

Lang and colleagues argued against routine preoperative laryngoscopic examination based on their findings in 302 patients undergoing nerve function assessment before thyroid surgery [11]. In their series, the prevalence of preoperative VFP was 2.3% and only one of the patients (0.4%) had not undergone previous thyroid surgery. Similarly in our study, the preoperative rate of VFP was 2.8%; however, in ten cases (1.1% of all patients), the etiology was idiopathic. In our practice, the reasons to perform preoperative laryngoscopy are (1) to register, in case of postoperative VFP, if it was indeed new and caused by the index procedure and (2), in case of preoperative VFP, to be aware of the risk of bilateral VFP after surgery. We had two patients with previous unilateral VFP who suffered from bilateral palsy after surgery. In both cases, the indication for the procedure was recurrent goiter and the preoperative VFP was caused by the primary procedure. Fortunately, in both cases, these new palsies were transient.

According to our results, the incidence of new VFP was 53 in 920 (5.8%) operations and 55 in 1296 (4.2%) nerves at risk. Compared to the results of previous studies in which routine laryngoscopy was used [4, 5], our complication rate was lower (5.8% against 7.6–13.9%). On the other hand, we observed much higher rate of permanent VFP than was demonstrated in these previous studies (3.2–3.4% against 0.9–1.4%). In our study, less than half of VFPs were transient in comparison with approximately 80%–90% in the other studies [4, 5]. These remarkable differences need to be analyzed focusing, for example, on patient’s characteristics, diagnoses, extent of surgical procedures, reporting standards, and hospital and surgeon volumes. In hospitals reporting low permanent VFP rates, operations were usually performed by one experienced surgical team. Our institution is a university clinic and teaching hospital, and consequently, many operations were done by surgeons in training. It is very important for any institution that performs thyroid surgery to acknowledge their RLN injury rates accurately, and the institutional risk must be discussed with the patient when weighing between surgery and surveillance. Patients need to be well informed of the rates of complications at the institution rather than giving the patients only vague estimates of the risks based on the literature. Although only a small proportion of patients with VFP were asymptomatic in our study (13%), postoperative VFP may be easily overlooked unless routine vocal fold examination is implemented to the surgical quality control regime [2].

The routine visualization of RLN is regarded as a gold standard for nerve injury prevention [12]. In our study, the rate of RLN identification was only 56%. Interestingly, we found no difference in VFP rates in cases where RLN was visualized compared to cases where RLN was not seen. Even so, the visualization of an intact RLN during operation correlated to normal VFP function postoperatively. The evidence on routine RLN visualization relies on cases series such as the recent publication by Dhillon and colleagues from Johns Hopkins Hospital which included 2527 nerves at risk with routine RLN evaluation by the operating surgeon in all cases; the reported rates of temporary VFP (2.9%) and permanent VFP (0.4%) per nerves at risk were significantly lower than in our study [13]. However, their study had a mixed cohort including parathyroid procedures, whereas redo procedures were not included; the postoperative VFP rate was 72/2153 nerves at risk (3.3%) in primary thyroid procedures compared to 55/1296 (4.2%) in our study with both primary and redo procedures. Furthermore, Dhillon’s study represents the vast experience of a single high-volume surgeon from one of the world’s most renowned hospitals, whereas our study is a “real-world” representation of small hospital outcomes. During our study period, intraoperative neuromonitoring was not used. Although it is widely recommended in special situations like redo surgery, according to two meta-analyses with 23,500 and 9000 pooled patients, routine use of neuromonitoring was not shown to decrease VFP rates [12, 14]. Considering the high rate of permanent VFP in our study, we modified our technique after this study to include neuromonitoring and we recommend routine identification of the RLN.

The present study confirmed that recurrent goiter is one of the most significant risk factors for postoperative VFP with nearly nine times higher risk than in all other indications combined. Previously, patients with benign recurrent goiter were shown to have 4.7 times higher risk of permanent VFP in a multi-institutional study with 16 448 procedures [6]. While this risk is well identified in the literature, it may still be underestimated in clinical work. The benefits of redo surgery must be carefully weighed against the high risk of RLN injury.

Other risk factors for RLN injury in this study were total thyroidectomy, sternotomy, lymph node dissection, hypocalcemia, and drain use. In total thyroidectomy, both RLNs were at risk, and therefore, the risk was at least twofold compared to hemithyroidectomy. Sternotomy, lymph node dissections, hypocalcemia, and drain use (in only 6% of operations) were all related to extensive surgery and, hence, to increased risk of RLN injury. The drain use itself has not been identified as a risk factor for VFP in a meta-analysis with 1927 patients [15]. In that study, drain was used more liberally, in 51% of patients. Our study was performed at a time when the use of energy devices (i.e., electronic sealing instruments) had been adapted to routine practice (used in 99% of all cases). The effect of energy devices in RLN injury rate could not be assessed because there were not enough cases for the control group; only a few patients were operated without any vessel sealing instrument.

Almost identically to a previous meta-analysis of 1798 patient, two-thirds of the patients with VFP in our study received voice therapy, whereas only 5.7% underwent injection laryngoplasty [16]. Early voice therapy in patients with unilateral VFP has been associated with better outcomes than late rehabilitation in one 11-year retrospective study with 171 patients [17]. Routine postoperative laryngoscopy enables early diagnosis and thus initiation of voice therapy without delay, which is also important in the prevention of aspiration-related problems. Furthermore, routine vocal fold screening gives direct feedback to the surgeon and may help avoid RLN injuries in the future.