Randomised controlled trial of in- versus out-patient management of benign hemithyroidectomy

Throughout the past century, the prevalence of outpatient (OPT) surgery has seen a significant rise. Advances in surgical and anaesthetic techniques have transformed procedures that once demanded days of inpatient (IPT) care into feasible OPT procedures [1]. This shift enhances resource utilisation while offering cost benefits. For instance, when thyroid procedures are conducted in an OPT setting, a documented 10-30% cost reduction is achieved [2]. However, paramount to these advantages is ensuring that patient safety is not compromised.

Historically, hemithyroidectomy was approached as an IPT procedure. This was largely due to the risk of potentially life-threatening post-operative haemorrhage, hypocalcaemia and paralysis of the recurrent laryngeal nerve (RLN) [3]. However, the occurrence of these risks is rare in hemithyroidectomies, rendering this procedure potentially safe as OPT [4].

Several studies have shown an increase in OPT thyroid surgery in recent decades, with the majority of studies agreeing that careful preoperative patient selection is crucial [5, 6]. In 2013, the American Thyroid Association suggested eligibility criteria for OPT thyroid surgery [5], and OPT thyroidectomy has become common in many US hospitals [7]. Europe, however, has embraced this trend more gradually [6]. Despite the global increase in OPT, many surgeons still abstain from performing hemithyroidectomy as an OPT procedure due to safety concerns. As healthcare systems and cultures differ, it is essential to perform trials investigating the applicability and safety of hemithyroidectomy in a European context.

We conducted a randomised clinical trial with patients undergoing hemithyroidectomy for benign thyroid disease, allocating them to either an IPT or OPT procedure. Our primary objective was to assess the rate of discharge failures in an OPT setting, defined as conversion to IPT or readmission within 24 hours, and to evaluate the complications leading to failure, especially post-operative haemorrhage. Furthermore, we aimed to evaluate patient-related outcome measures at the level of post-operative pain, nausea and, finally, acceptance of and satisfaction with the IPT or OPT procedures.

METHODS

This randomised, parallel group, non-inferiority study was approved by the local regional ethical committee (#SJ-495) and registered with clinicaltrials.gov (NCT02891252). The inclusion and exclusion criteria are listed in Table 1. Hemithyroidectomies were performed by seven experienced thyroid surgeons using Kocher's incision, Nerve Integrity Monitor and standardised closure techniques. Choice of surgical instruments, including LigaSure and Surgicel, was at the surgeon’s discretion. Drains were placed if necessary, converting OPT to IPT.

Preoperative medication included oral celecoxib 200 mg and paracetamol 1,000 mg. Anaesthesia was standardised using propofol and remifentanil infusions. Dexamethasone 16 mg, administered intravenously (i.v), was used for both pain and as anti-emetic prophylaxis, together with ondansetron i.v. 4 mg. Post-operatively, patients were transferred to the post-anaesthesia care unit (PACU) and received i.v. oxycodone as needed for pain. In the ward, standard observation and pain relief were provided. Discharge followed the department’s OPT criteria, with the doctor in charge assessing each patient. All patients were observed for a minimum of six hours and had a fibreoptic laryngoscopy prior to discharge.

Primary and secondary outcomes are described in Table 1. Patient-related outcome measures (PROM) were registered by the patients using a numeric rating scale (NRS 0-10), including pain, nausea, sleep and fatigue. Vomiting episodes and opioid administration were noted at the PACU. Additional PROMs included NRS-rated satisfaction and days until patients could resume normal activity (questionnaire in the supplementary material). Complications were registered retrospectively from the patient’s medical records. OPT were defined as discharge on the day of surgery. Failures were defined as OPT converted into IPT or re-admission within 24 hours. Complications were categorised as either minor or major.

Preoperative and perioperative data were recorded, including gender, age, indication for surgery, vocal cord function, anticoagulation use, thyroid-stimulating hormone (TSH), thyroxine (T₄), gland weight, perioperative bleeding, LigaSure use, Surgicel use, surgery time and anaesthesia time.

The power calculation was based on a 20% limit of non-inferiority between the IPT and OPT groups .A total of 86 patients (43 in each group) were required to ensure sufficient power. Block randomisation with a 4:1 female-to-male ratio was used. Randomisation lists were kept in sealed opaque envelopes. Neither patients nor surgeons were blinded during the study. Data were analysed using SPSS version 28.0, with χ² test for dichotomous parameters and Student’s t-test for continuous parameters. Pearson’s correlation was applied across all parameters, considering r values < 0.40 as low or no correlation. Both IPT and OPT groups were analysed as intention to treat.

Trial registration: ClinicalTrials.gov Identifier: NCT02891252

RESULTS

Between May 2016 and November 2021, 97 patients were included (80 women and 17 men); 46 in the IPT and 51 in the OPT group (Table 2). Indications for surgery were most often compressive symptoms (48%) and, following the Danish guidelines, if indeterminate cytology or nodule growth was present (16%) (Table 2). No strong correlations were found using Pearson’s correlation. No differences were observed in patient demographics or perioperative characteristics (Table 2).

Outcome

Overall, complications were reported for 11% of the IPT group and 29% of the OPT group (Table 3). We observed three major complications; all were RLN injuries; one definitive and two transient. In the OPT group, 14 patients (27.5%) were converted to the IPT procedure and therefore registered as failures. The majority of failures were due to nausea (36%) and neck swelling (29%). The rate of failures decreased as the study progressed (data not shown). Specifically, 64% of all failures occurred among the first 40 patients included, and 79% of all failures among the first 60 patients. After the inclusion of 80 patients, no additional failures were recorded. None of the patients in either group required reoperation or readmission.

An unsuspected diagnosis of thyroid carcinoma was found in 12% (data not shown).

Overall, 72 patients (74.2%) handed in a complete PROM questionnaire; 69.6% (32/46) in the IPT and 78.4% (40/51) in the OPT group of whom six were failures.

No significant difference was observed in levels of nausea and pain in the two groups upon leaving the PACU (Table 4). However, at 8 PM on post-operative day 0 (POD0) and at 8 AM at POD1, the OPT group reported significantly less nausea than the IPT group. Additionally, the OPT group reported better sleep the first night. No other differences were observed.

The OPT group reported significantly fewer days until resuming normal daily activity than the IPT group, 5.9 versus 9.4 days, respectively (Table 4). Both groups reported an equally high level of satisfaction.

DISCUSSION

This randomised study evaluated the feasibility of OPT hemithyroidectomy for benign conditions. A total of 27.5% of the OPT patients in this group failed to be discharged on the same day. However, no severe haemorrhages or re-operations occurred. The OPT group experienced less post-operative nausea, better first-night sleep and fewer days until resuming normal activity. Patient satisfaction was similar in both groups with no single parameter correlating with complications or discharge failures.

Even though thyroid surgery is increasingly performed as OPT, most studies are retrospective [2, 8] and heterogeneous [2, 3]. By including only benign hemithyroidectomies, this study achieved a homogeneous cohort with a low risk of complications. Careful patient selection is universally agreed upon as essential for OPT thyroid surgery [5, 6], with factors such as proximity to skilled facilities, absence of major comorbidities and appropriate social settings considered [5, 6].

Changing routines may be challenging and are influenced by both cultural norms and distinctive geographical challenges. In Denmark, widespread rural regions and numerous islands, some lacking direct bridge connections, often translate into longer hospital commutes. This can delay crucial post-operative care, especially when winding roads or ferry routes are involved. Such challenges are less pronounced in urban areas with nearby healthcare facilities, making OPT more straightforward. In contrast, rural regions face inherent complexities due to these logistical challenges. Moreover, while OPT is becoming more common globally, its adoption in Denmark may potentially be influenced not just by habit and culture but also by these geographical and infrastructural considerations. In this study, resistance to discharging OPT patients decreased over time, indicating a learning curve for providers. Some neck swelling and mild complications like nausea are expected and may not require hospitalisation. Failure to discharge on the same day is inevitable in some cases, with observed failure rates ranging from 9% to 42% in studies involving only hemithyroidectomies [6]. It must be noted that some institutions discharge patients with drainage. At our department, we do not.

In this study, the average time to discharge for OPT was ten hours. Our protocol mandates a minimum six-hour interval between surgery and discharge, aligning with the practices of several other centres [5, 9-11]. Given the logistical concerns and considering patients’ welfare, nighttime discharges are not feasible. To address this, patients participating in the project were typically scheduled as the first surgeries of the day, ensuring that they could be evaluated for discharge during the evening rounds.

Post-operative nausea and vomiting is a common challenge [9]. In this study, 36% of failures were due to nausea despite anti-emetic prophylaxis given as standard treatment. No significant difference in morphine use was found between OPT failures and the remaining OPT group. OPT patients reported less nausea than the IPT group. Neck swelling resulted in 28.6% (four out of 14) failures in our study. Neck swelling was assessed clinically without the use of ultrasound or other measures and decreased in the course of the study period. The doctor in charge was not instructed to evaluate both OPT and IPT patients equally regarding complications, and the higher complication rate in the OPT group may potentially be explained by not reporting minor problems, such as neck swelling, in our IPT patients.

No post-operative haemorrhages occurred, and a recent systematic review supported the low risk in OPT hemithyroidectomies [12]. Surgery time did not differ between groups, suggesting that equally careful haemostasis was employed in the OPT and IPT groups. Patient satisfaction was generally high in the OPT group. While other studies have shown similar results [1, 6, 9], very few have addressed patient satisfaction, and none employed comparable measures.

The strength of this study lies in its randomisation, minimising confounding and selection, allocation and performance bias. However, weaknesses include the 74% return rate for PROM questionnaires, potentially introducing selection bias. The study deliberately employed non-validated questionnaires to enhance patient compliance. Our primary objective was to compare differences between the two groups rather than to obtain absolute scores for broader comparisons with other populations or studies. The study was not blinded and objective criteria for assessing swelling were lacking. A further limitation of the study is that we did not track the number of patients who chose OPT surgery versus those who declined participation. Additionally, we did not record reasons why patients were deemed ineligible. Such data would have offered valuable insights.

CONCLUSIONS

This study did not demonstrate non-inferiority of OPT compared to IPT hemithyroidectomy. However, no serious complications or reoperations were observed. The failure rate was initially higher, suggesting a transgressive discharge process. Patient satisfaction was high, and no re-admittances were recorded. We suggest that hemithyroidectomy for benign indications is a feasible procedure in an OPT setting for a selected group of patients, given the mild complications and overall satisfaction.