Thyroid surgery for Graves’ disease and Graves’ ophthalmopathy

Abstract

Background

Graves' disease is an autoimmune disease caused by the production of auto‐antibodies against the thyroid‐stimulating hormone receptor, which stimulates follicular cell production of thyroid hormone. It is the commonest cause of hyperthyroidism and may cause considerable morbidity with increased risk of cardiovascular and respiratory adverse events. Five per cent of people with Graves' disease develop moderate to severe Graves' ophthalmopathy. Thyroid surgery for Graves' disease commonly falls into one of three categories: 1) total thyroidectomy, which aims to achieve complete macroscopic removal of thyroid tissue; 2) bilateral subtotal thyroidectomy, in which bilateral thyroid remnants are left; and 3) unilateral total and contralateral subtotal thyroidectomy, or the Dunhill procedure. Recent American Thyroid Association guidelines on treatment of Graves' hyperthyroidism emphasised the role of surgery as one of the first‐line treatments. Total thyroidectomy removes target tissue for the thyroid‐stimulating hormone receptor antibody. It controls hyperthyroidism at the cost of lifelong thyroxine replacement. Subtotal thyroidectomy leaves a thyroid remnant and may be less likely to lead to complications, however a higher rate of recurrent hyperthyroidism is expected and revision surgery would be challenging. The choice of the thyroidectomy technique is currently largely a matter of surgeon preference, and a systematic review of the evidence base is required to determine which option offers the best outcomes for patients.

Objectives

To assess the optimal surgical technique for Graves' disease and Graves' ophthalmopathy.

Search methods

We searched the Cochrane Library, MEDLINE and PubMed, EMBASE, ClinicalTrials.gov, and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP). The date of the last search was June 2015 for all databases. We did not apply any language restrictions.

Selection criteria

Only randomised controlled trials (RCTs) involving participants with a diagnosis of Graves' disease based on clinical features and biochemical findings of hyperthyroidism were eligible for inclusion. Trials had to directly compare at least two surgical techniques of thyroidectomy. There was no age limit to study inclusion.

Data collection and analysis

Two review authors independently extracted and cross‐checked the data for analysis, evaluation of risk of bias and establishment of 'Summary of findings' tables using the GRADE instrument. The senior review authors reviewed the data and reconciled disagreements.

Main results

We included five RCTs with a total of 886 participants; 172 were randomised to total thyroidectomy, 383 were randomised to bilateral subtotal thyroidectomy, 309 were randomised to the Dunhill procedure and 22 were randomised to either bilateral subtotal thyroidectomy or the Dunhill procedure. Follow‐up ranged between six months and six years. One trial had three comparison arms. All five trials were conducted in university hospitals or tertiary referral centres for thyroid disease. All thyroidectomies were performed by experienced surgeons. The overall quality of the evidence ranged from low to moderate. In all trials, blinding procedures were insufficiently described. Outcome assessment for objective outcomes was blinded in one trial. Surgeons were not blinded in any of the trials. One trial blinded participants. Attrition bias was a substantial problem in one trial, with 35% losses to follow‐up. In one trial the analysis was not carried out on an intention‐to‐treat basis.

Total thyroidectomy was more effective than subtotal thyroidectomy techniques (both bilateral subtotal thyroidectomy and the Dunhill procedure) at preventing recurrent hyperthyroidism in 0/150 versus 11/200 participants (OR 0.14 (95% CI 0.04 to 0.46); P = 0.001; 2 trials; moderate quality evidence). Total thyroidectomy was also more effective than bilateral subtotal thyroidectomy at preventing recurrent hyperthyroidism in 0/150 versus 10/150 participants (odds ratio (OR) 0.13 (95% confidence interval (CI) 0.04 to 0.44); P = 0.001; 2 trials; moderate quality evidence). Compared to bilateral subtotal thyroidectomy, the Dunhill procedure was more likely to prevent recurrent hyperthyroidism in 20/283 versus 8/309 participants (OR 2.73 (95% CI 1.28 to 5.85); P = 0.01; 3 trials; low quality evidence). Total thyroidectomy compared with subtotal thyroidectomy conferred a greater risk of permanent hypocalcaemia/hypoparathyroidism in 8/172 versus 3/221 participants (OR 4.79 (95% CI 1.36 to 16.83); P = 0.01; 3 trials; low quality evidence). Effects of the various surgical techniques on permanent recurrent laryngeal nerve palsy and regression of Graves' ophthalmopathy were neutral. One death was reported in one study in year three of follow‐up. No study investigated health‐related quality of life or socioeconomic effects.

Authors' conclusions

Total thyroidectomy is more effective than subtotal thyroidectomy (both bilateral subtotal thyroidectomy and the Dunhill procedure) at preventing recurrent hyperthyroidism in Graves' disease. The type of surgery performed does not affect regression of Graves’ ophthalmopathy. There was some evidence that total thyroidectomy compared with subtotal thyroidectomy conferred a greater risk of permanent hypocalcaemia/hypoparathyroidism, which however, was not seen in comparison with bilateral subtotal thyroidectomy. Permanent recurrent laryngeal nerve palsy did not seem to be affected by type of thyroidectomy. Health‐related quality of life as a patient‐important outcome measure should form a core determinant of any future trial on the effects of thyroid surgery for Graves' disease.

Plain language summary

Thyroid surgery for Graves' disease and Graves' opthalmopathy

Review question

The aim of this review was to compare different surgical techniques for treatment of Graves' disease. We wanted to address whether surgically removing the whole thyroid (total thyroidectomy) gland is better than removing most of the gland (subtotal thyroidectomy) at controlling increased activity of the thyroid (hyperthyroidism) and eye symptoms associated with Graves' disease.

Background

Graves' disease is a common condition affecting the thyroid gland, which has an important role in controlling the body's metabolism. The body's own immune system attacks the thyroid gland and causes it to work much harder. This overactivity can lead to increased risk of heart attacks and strokes if left untreated for a long period of time. Graves' disease also causes protrusion of the eye and may limit eye movement (Graves' ophthalmopathy). Graves' disease is generally treated by medication alone or sometimes with radiation or surgery. In recent years, there has been increasing interest in surgery for Graves' disease and other thyroid problems, as results of surgery and health‐related quality of life outcomes appear promising. However, there is controversy over which surgical technique is best, as total thyroidectomy may give better control of disease long term, compared with subtotal thyroidectomy, but carry a greater risk of postoperative complications, such as damage to the so‐called recurrent laryngeal nerve (causing vocal cord dysfunction).

Study characteristics

We identified five randomised controlled studies that were relevant to the review question that directly compared different thyroidectomy surgical techniques. A total of 886 participants were included in the studies and follow‐up was between six months and six years. All five studies were conducted in university hospitals or tertiary referral centres for thyroid disease. The countries involved were Germany, Sweden, Poland and Taiwan. No study was performed in an outpatient setting.

Key results

Total thyroidectomy is likely to give better control of hyperthyroidism compared to subtotal thyroidectomy (8 or 9 people out of 1000 will experience recurrence of hyperthyroidism after total thyroidectomy compared with 55 to 67 people out of 1000 after subtotal thyroidectomy). There is some evidence to suggest permanently low blood calcium (hypoparathyroidism) is more likely in total thyroidectomy compared to subtotal thyroidectomy (59 people out of 1000 after total thyroidectomy compared with 14 people out of 1000 after subtotal thyroidectomy will experience this side effect). Eye symptoms as well as permanent damage to the recurrent laryngeal nerve appear to be unrelated to the choice of surgical technique. Deaths were rarely reported in the included studies. No study evaluated health‐related quality of life or socioeconomic effects (such as absence from work).

Quality of the evidence

The studies in this review suffer from a lack of high quality evidence. Several risk of bias problems were visible such as objective measurements of outcomes without knowing the allocation of participants to the intervention groups (total or subtotal thyroidectomy) and difficulties in finding out if all planned outcomes in studies were really reported in the publications of these studies. Also, rates of outcomes (event rates) were rather low, making our estimation of the differences between the surgical techniques imprecise. Finally, all thyroidectomies were performed by experienced surgeons and it is unknown whether the described results will be the same in other settings.

Currentness of evidence

This plain language summary is current as of 22 June 2015.

Summary of findings

Summary of findings for the main comparison. Total thyroidectomy compared to subtotal thyroidectomy for Graves disease and Graves' opthalmopathy.

Total thyroidectomy compared to subtotal thyroidectomy for Graves disease/Graves opthalmopathy
Patient: patients with Graves' disease/Graves' opthalmopathy Settings: university hospitals in Poland, Germany and Sweden Intervention: total thyroidectomy Comparison: subtotal thyroidectomy
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed riska	Corresponding risk
	Subtotal thyroidectomy	Total thyroidectomy
Rate of recurrent hyperthyroidism [measured by thyroid function tests] Follow‐up: 6 months and 6 years	55 per 1000	8 per 1000 (2 to 26)	OR 0.14 (0.04 to 0.46)	350 (2)	⊕⊕⊕⊝ moderateb	‐
Regression of Graves' ophthalmopathy [ophthalmology assessment] Follow‐up: 6 months and 6 years	782 per 1000	805 per 1000 (697 to 882)	OR 1.15 (0.64 to 2.08)	229 (2)	⊕⊕⊝⊝ lowc	‐
Permanent recurrent laryngeal nerve palsy [patient reported, confirmation by laryngoscopy] Follow‐up: 6 months to 6 years	18 per 1000	26 per 1000 (7 to 93)	OR 1.45 (0.38 to 5.59)	393 (3)	⊕⊕⊝⊝ lowd	‐
Permanent hypocalcaemia/hypoparathyroidism [measured by serum calcium levels] Follow‐up: 6 months to 6 years	14 per 1000	59 per 1000 (16 to 190)	OR 4.79 (1.36 to 16.83)	393 (3)	⊕⊕⊝⊝ lowd	‐
All‐cause mortality Follow‐up: 6 months to 6 years	See comment	See comment	Not estimable	350 (1)	See comment	Only one death reported in total thyroidectomy group
Health‐related quality of life	See comment	See comment	Not estimable	See comment	See comment	Not investigated
Socioeconomic effects	See comment	See comment	Not estimable	See comment	See comment	Not investigated
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; OR: Odds ratio; RR: Risk ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

^aAssumed risk was derived from the event rates in the intervention comparator groups.
 ^bDowngraded by one level because of high risk of performance bias in both trials and reporting bias in one trial.
 ^cDowngraded by two levels because of high risk of performance bias in both trials and imprecision.
 ^dDowngraded by two levels because of high risk of performance bias in all trials, imprecision and low event rates.

Summary of findings 2. Total thyroidectomy compared to bilateral subtotal thyroidectomy for Graves disease and Graves' opthalmopathy.

Total thyroidectomy compared to bilateral subtotal thyroidectomy for Graves disease/Graves' opthalmopathy
Patient: patients with Graves' disease/Graves' opthalmopathy Settings: university hospitals in Germany and Poland Intervention: total thyroidectomy Comparison: bilateral subtotal thyroidectomy
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed riska	Corresponding risk
	Bilateral subtotal thyroidectomy	Total thyroidectomy
R ate of recurrent hyperthyroidism [measured by thyroid function tests] Follow‐up: 6 months and 6 years	67 per 1000	9 per 1000 (3 to 30)	OR 0.13 (0.04 to 0.44)	300 (2)	⊕⊕⊕⊝ moderateb	‐
Regression of Graves' ophthalmopathy [ophthalmology assessment] Follow‐up: 6 months and 6 years	791 per 1000	818 per 1000 (707 to 893)	OR 1.19 (0.64 to 2.21)	260 (2)	⊕⊕⊝⊝ lowc	‐
Permanent recurrent laryngeal nerve palsy [patient reported, confirmation by laryngoscopy] Follow‐up: 6 months and 6 years	13 per 1000	13 per 1000 (2 to 88)	OR 1.00 (0.14 to 7.15)	300 (2)	⊕⊕⊝⊝ lowd	‐
Permanent hypocalcaemia/hypoparathyroidism [measured by serum calcium levels] Follow‐up: 6 months and 6 years	20 per 1000	40 per 1000 (11 to 138)	OR 2.04 (0.53 to 7.85)	300 (2)	⊕⊕⊝⊝ lowd	‐
All‐cause mortality Follow‐up: 6 months and 6 years	See comment	See comment	Not estimable	300 (2)	See comment	Only one death reported in total thyroidectomy group
Health‐related quality of life	See comment	See comment	Not estimable	See comment	See comment	Not investigated
Socioeconomic effects	See comment	See comment	Not estimable	See comment	See comment	Not investigated
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; OR: Odds ratio; RR: Risk ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

^aAssumed risk was derived from the event rates in the intervention comparator groups.
 ^bDowngraded by one level because of high risk of performance bias in both trials.
 ^cDowngraded by two levels because of high risk of performance bias in both trials and imprecision.
 ^dDowngraded by two levels because of high risk of performance bias in both trials, imprecision and low event rates.

Summary of findings 3. Bilateral subtotal thyroidectomy compared to the Dunhill procedure for Graves disease and Graves' opthalmopathy.

Bilateral subtotal thyroidectomy compared to the Dunhill procedure for Graves disease/Graves' opthalmopathy
Patient: patients with Graves disease/Graves' opthalmopathy Settings: university hospitals in Taiwan and Germany Intervention: bilateral subtotal thyroidectomy Comparison: Dunhill procedure
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed riska	Corresponding risk
	Dunhill procedureb	Bilateral subtotal thyroidectomy
R ate of recurrent hyperthyroidism [measured by thyroid function tests] Follow‐up: 6 months to 6 years	26 per 1000	68 per 1000 (33 to 135)	OR 2.73 (1.28 to 5.85)	592 (3)	⊕⊕⊝⊝ lowc	‐
Regression of Graves' ophthalmopathy [ophthalmology assessment] Follow‐up: 6 months and 6 years	689 per 1000	633 per 1000 (415 to 810)	OR 0.78 (0.32 to 1.92)	90 (2)	⊕⊕⊝⊝ lowd	‐
Permanent recurrent laryngeal nerve palsy [patient reported, confirmation by laryngoscopy] Follow‐up: 6 months to 6 years	3 per 1000	0 per 1000 (0 to 22)	OR 0.14 (0.00 to 6.82)	592 (3)	⊕⊕⊝⊝ lowc	‐
Permanent hypocalcaemia/hypoparathyroidism [measured by serum calcium levels] Follow‐up: 6 months to 6 years	10 per 1000	18 per 1000 (5 to 70)	OR 1.90 (0.47 to 7.71)	592 (3)	⊕⊕⊝⊝ lowc	‐
All‐cause mortality Follow‐up: 6 months to 6 years	See comment	See comment	Not estimable	642 (3)	See comment	No deaths reported in any study
Health‐related quality of life	See comment	See comment	Not estimable	See comment	See comment	Not investigated
Socioeconomic effects	See comment	See comment	Not estimable	See comment	See comment	Not investigated
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; OR: Odds ratio; RR: Risk ratio
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

^aAssumed risk was derived from the event rates in the intervention comparator groups
 ^bDunhill procedure: unilateral total and contralateral subtotal thyroidectomy
 ^cDowngraded by two levels because of high risk of performance bias in all trials, attrition bias in one of three trials, imprecision and low event rates
 ^dDowngraded by two levels because of high risk of performance bias in both trials, attrition bias in one of two trials, imprecision and low number of participants

Background

Description of the condition

Graves' disease accounts for 50% to 80% of cases of hyperthyroidism and has a population prevalence of 0.5% to 2%. It is more common in women, with a female to male ratio of 5:1 (Genovese 2012; Ponto 2012). Five per cent of patients with Graves' disease develop moderate to severe Graves' ophthalmopathy, with older age groups (40 to 80 years) at higher risk of developing eye signs (Laurberg 2012).

Graves' disease may be diagnosed by evidence of extra‐thyroidal manifestations on examination and biochemical hyperthyroidism (suppressed thyroid‐stimulating hormone, high tri‐iodothyronine (T3), thyroxine (T4) or both). If the diagnosis is uncertain, a radionuclide uptake test may be performed which shows diffusely increased uptake throughout the thyroid. Assays for the thyroid‐stimulating hormone receptor antibody have 98% sensitivity and 100% specificity for Graves' disease, but their use in clinical practice remains variable (Lalwani 2012).

Hyperthyroid states are associated with significant cardiovascular and respiratory adverse events such as atrial fibrillation, congestive cardiac failure, hypercoagulability and stroke. A recent meta‐analysis suggested a 20% increase in mortality in patients diagnosed with hyperthyroidism (Brandt 2011).

Description of the intervention

Thyroid surgery for Graves' disease commonly falls into one of three categories: 1) total thyroidectomy, which aims to achieve complete macroscopic removal of thyroid tissue; 2) bilateral subtotal thyroidectomy, in which bilateral thyroid remnants are left; and 3) unilateral total and contralateral subtotal thyroidectomy, or the Dunhill procedure. In subtotal thyroidectomy techniques, 1 mL to 4 mL of thyroid remnant per lobe is recommended, to reduce the risk of recurrent hyperthyroidism (Chi 2005; Hermann 1998).

Adverse effects of the intervention

Thyroidectomy for benign disease carries significantly less risk than for thyroid malignancies. Two recent large retrospective case series for total thyroidectomy in benign disease (Bellantone 2002;Efremidou 2009) reported the risk of permanent recurrent laryngeal nerve palsy as 0% to 0.4%, and temporary recurrent laryngeal nerve palsy at 1.3%. Temporary hypocalcaemia was seen in 7.3% of patients, and permanent hypocalcaemia was seen in 0.3% to 3.4%. Haemorrhage requiring repeat surgery was reported to be in the region of 0.2% to 1.5%. The trend was for a decrease in complications over recent years, suggesting surgeon experience and caseload may be factors in complication rates. Subtotal thyroidectomy has been reported to give lower rates of temporary recurrent laryngeal nerve palsy and hypocalcaemia, but outcomes are similar in the long term (Barczynski 2011).

How the intervention might work

Total thyroidectomy removes target tissue for the thyroid‐stimulating hormone receptor antibody. It controls hyperthyroidism at the cost of lifelong thyroxine replacement. Subtotal thyroidectomy leaves a thyroid remnant and aims to achieve euthyroidism without the need for thyroxine replacement, however a higher rate of recurrent hyperthyroidism is expected.

Graves' ophthalmopathy is thought to be due to common antigens shared between the thyroid and orbit that are both targeted by the thyroid‐stimulating hormone receptor antibody and a G‐protein G2sAb. Total thyroidectomy has been shown to be associated with regression of Graves' opthalmopathy, associated with a concurrent decrease in the thyroid‐stimulating hormone receptor antibody and G2sAb postsurgery (De Bellis 2012; Leo 2012).

Why it is important to do this review

Thyroidectomy for Graves' disease may be indicated in patients who have persistent hyperthyroidism after a trial of anti‐thyroid medication; when a rapid return to euthyroidism is desired; or when radioactive iodine treatment is contraindicated (e.g. in pregnancy).

In recent years we have seen a resurgence of interest in surgical management of Graves' disease. The American Thyroid Association's Hyperthyroidism Management Guidelines 2011 named surgery as one of three first‐line treatments for hyperthyroidism associated with Graves' disease alongside radioactive iodine and anti‐thyroid drugs (ATA 2011). In studies conducted in patients who have persistent hyperthyroidism after medical treatment, total thyroidectomy is shown to be more cost‐effective than radioactive iodine or lifelong anti‐thyroid medication (In 2009; Zanocco 2012). Health‐related quality of life after surgical ablation of the thyroid has been shown to be comparable to the general population (Al‐Adhami 2012). Morbidity associated with thyroidectomy has been shown to be less than previously thought (Bellantone 2002; Efremidou 2009).

Choice of thyroidectomy technique is largely a matter of surgeon preference, and a systematic review of the evidence base is required to determine which option offers the best outcome for patients. A previous meta‐analysis on this subject was carried out in 2000 (Palit 2000), based on a MEDLINE search. Since that time, data from several new randomised controlled trials (RCTs) have been published, which demonstrates considerable interest in this topic.

Objectives

To assess the optimal surgical technique for Graves' disease and Graves' ophthalmopathy.

Methods

Criteria for considering studies for this review

Types of studies

We included only RCTs.

Types of participants

All patient age groups receiving surgical treatment for Graves' disease.

Diagnostic criteria

Our criteria for Graves' disease were clinical examination and suppressed thyroid‐stimulating hormone and elevated free thyroxine (T4), free tri‐iodothyronine (T3), or both. Additional diagnostic tests such as radionuclide uptake and thyroid‐stimulating hormone receptor antibody)/thyrotropin binding inhibiting immunoglobulins were noted where performed, but not considered essential for study inclusion.

Types of interventions

The primary aim of our evaluation was to compare total thyroidectomy with subtotal thyroidectomy for Graves' disease and Graves' opthalmopathy. However, to provide maximum information we also included all types of thyroidectomy for these conditions such as bilateral thyroidectomy versus unilateral total and contralateral subtotal thyroidectomy (the Dunhill procedure).

We excluded studies where only one trial arm contained a surgical intervention, or where surgical interventions were compared to medical interventions.

Types of outcome measures

Primary outcomes

• Rate of recurrent hyperthyroidism.

• Adverse events.

Secondary outcomes

• Regression of Graves' ophthalmopathy.

• All‐cause mortality.

• Socioeconomic effects.

• Health‐related quality of life.

Method and timing of outcome measurement

• Rate of recurrent hyperthyroidism: biochemically confirmed elevation of T3/T4 with postoperatively suppressed thyroid‐stimulating hormone and measured at more than three months postsurgery.

• Adverse events such as: permanent recurrent laryngeal nerve palsy defined as persistent recurrent laryngeal nerve palsy either objectively measured by laryngoscopy or reported by the patient, and persisting more than six months postsurgery. Permanent hypocalcaemia/hypoparathyroidism is measured as persistent corrected serum calcium below the reference range as per local laboratory assays more than six months postsurgery. Postoperative bleeding, defined as bleeding sufficiently severe to warrant intervention such as suture removal or return to theatre and measured at less than 28 days postsurgery.

• Regression of Graves' ophthalmopathy: a clinically significant improvement in Graves' ophthalmopathy using a validated scoring system as reported by the authors of the study and measured at more than six months postsurgery.

• All‐cause mortality: defined as death from any cause measured at any time during the study.

• Health‐related quality of life: measured with a validated instrument such as SF‐36, WHOQOL and measured at any time during the study.

• Socioeconomic effects: defined as any qualitative or quantitative socioeconomic measure such as cost‐benefit analyses and measured at any time during the study.

Summary of findings' table

We present 'Summary of findings' tables using the following outcomes listed according to priority.

1. Rate of recurrent hyperthyroidism.

2. Regression of Graves' ophthalmopathy.

3. Permanent recurrent laryngeal nerve palsy.

4. Permanent hypocalcaemia/hypoparathyroidism.

5. All‐cause mortality.

6. Health‐related quality of life.

7. Socioeconomic effects.

Search methods for identification of studies

Electronic searches

We searched the following sources from inception of each database to the specified date and placed no restrictions on the language of publication.

• Cochrane Library (22 June 2015).

  • Cochrane Central Register of Controlled Trials (Issue 5 of 12, May 2015).

  • Database of Abstracts of Reviews of Effect (Issue 2 of 4, April 2015).

  • Health Technology Assessment Database (Issue 2 of 4, April 2015).

• MEDLINE(R) In‐Process & Other Non‐Indexed Citations and Ovid MEDLINE(R) 1946 to Present (22 June 2015) and PubMed (subsets not available on Ovid, 22 June 2015).

• EMBASE 1974 to 2015 Week 25 (22 June 2015).

• ClinicalTrials.gov. (22 June 2015).

• ICTRP Search Portal (http://apps.who.int/trialsearch/, 22 June 2015).

  • Australian New Zealand Clinical Trials Registry (15 June 2015).

  • Chinese Clinical Trial Registry (15 June 2015).

  • ClinicalTrials.gov (15 June 2015).

  • EU Clinical Trials Register (15 June 2015).

  • ISRCTN (15 June 2015).

  • The Netherlands National Trial Register (15 June 2015).

  • Brazilian Clinical Trials Registry (26 May 2015).

  • Clinical Trials Registry ‐ India (26 May 2015).

  • Clinical Research Information Service ‐ Republic of Korea (26 May 2015).

  • Cuban Public Registry of Clinical Trials (26 May 2015).

  • German Clinical Trials Register (1 June 2015).

  • Iranian Registry of Clinical Trials (26 May 2015).

  • Japan Primary Registries Network (26 May 2015).

  • Pan African Clinical Trial Registry (26 May 2015).

  • Sri Lanka Clinical Trials Registry (1 June 2015).

  • Thai Clinical Trials Register (26 May 2015).

We continuously applied a MEDLINE (via Ovid SP) email alert service established by the Cochrane Metabolic and Endocrine Disorders (CMED) Group to identify newly published trials using the same search strategy as described for MEDLINE (for details on search strategies, see Appendix 1).

In case we detected new studies for inclusion we would have evaluated these and incorporated findings in our review before submission of the final review draft (Beller 2013). If additional keywords of relevance were detected during any of the electronic or other searches, we would have modified the electronic search strategies to incorporate these terms.

Searching other resources

We tried to identify other potentially eligible trials or ancillary publications by searching the reference lists of retrieved included trials, (systematic) reviews, meta‐analyses and health technology assessment reports.

Data collection and analysis

Selection of studies

Two review authors (ZWL, LM) independently scanned the abstract, title or both sections of every record retrieved to determine the studies to be assessed further. We investigated all potentially relevant articles as full text. A third party (BF, PJ or KC) resolved any differences in opinion. If resolving disagreements was not possible, we added the article to those 'awaiting assessment' and we contacted authors for clarification. We attach a PRISMA (preferred reporting items for systematic reviews and meta‐analyses) flow‐chart of study selection (Liberati 2009).

Data extraction and management

For studies that fulfilled inclusion criteria, two review authors (ZWL, LM) independently summarised relevant population and intervention characteristics using standard data extraction templates (for details see Characteristics of included studies; Table 4; Appendix 2; Appendix 3; Appendix 4; Appendix 5; Appendix 6; Appendix 7; Appendix 8; Appendix 9; Appendix 10) with any disagreements resolved by discussion, or, if required, by a third party (BF, PJ or KC).

1. Overview of study populations.

Characteristic	Intervention(s) and comparator(s)	Sample sizea	Screened/eligible [N]	Randomised [N]	ITT [N]	Finishing study [N]	Randomised finishing study [%]	Follow‐up
(1) Barczynski 2012	I: total thyroidectomy	(1) To detect an estimated absolute difference in the rate of recurrent hyperthyroidism of 5%, it was calculated that 71 participants would be required in each treatment arm to give the study a power of 80% (2) To detect an estimated difference in ophthalmopathy progression rate of 8%, 83 participants were required in each treatment arm to give the study a power of 80% Anticipating a 15% loss to follow‐up, a total of 200 participants were included	298	100	100	95	95.0	5 years
	C: bilateral subtotal thyroidectomy			100	100	96	96.0
	total:			200	200	191	95.5
(2) Witte 2000	I: total thyroidectomy	‐	‐	50	50	‐	‐	6‐36 (58) months
	C1: bilateral subtotal thyroidectomyb			50	50	‐	‐
	C2: Dunhill procedureb			50	50	‐	‐
	total:			150	150	‐	‐
(3) Chi 2005	I: bilateral subtotal thyroidectomy	‐	‐	166	‐	166	100	3‐26.4 months
	C: Dunhill procedure			174	‐	174	100
	total:			340		340	100
(4) Jarhult 2005	I: total thyroidectomy	‐	‐	22	‐	21	100	3 years
	C: subtotal thyroidectomy			22	‐	21	95.5
	total:			44		43	97.7
(5) Muller 2001	I: bilateral subtotal thyroidectomy	‐	‐	67	‐	46	68.7	6 years
	C: Dunhill procedure			85	‐	53	62.4
	total:			152	‐	99	65.1
Grand total	All interventions			405		368
	All comparators			481		344
	All interventions and comparators			886		712

^aAccording to power calculation in study publication or report.
 ^bSubtotal thyroid resection was performed in 103 patients with an attempt to leave less than 4 mL total remnant

"‐" denotes not reported.

C: comparator; I: intervention; ITT: intention‐to‐treat; N/A: not applicable.

We provide information about registered studies in the table Characteristics of ongoing studies and in the appendix 'Matrix of study endpoints (publications and trial documents'; Appendix 5).

We sent an email request to study authors of included studies to enquire whether they were willing to answer questions regarding their trials. Appendix 11 shows the results of this survey. Thereafter, we sought relevant missing information on the trial from the study authors of the article, if required.

Dealing with duplicate and companion publications

In the event of duplicate publications, companion documents or multiple reports of a primary study, we planned to maximise yield of information by collating all available data. In case of doubt, the publication reporting the longest follow‐up associated with our primary or secondary outcomes would have obtained priority.

Assessment of risk of bias in included studies

Two review authors (ZWL, LM) assessed each trial independently. Possible disagreements were resolved by consensus, or with consultation of a third party (BF, PJ or KC). In cases of disagreement, seniors authors were consulted and a consensus reached.

We assessed risk of bias using the Cochrane Collaboration’s tool (Higgins 2011a; Higgins 2011b). We used the following bias criteria.

• Random sequence generation (selection bias).

• Allocation concealment (selection bias).

• Blinding (performance bias and detection bias), separated for blinding of participants and personnel, and blinding of outcome assessment.

• Incomplete outcome data (attrition bias).

• Selective reporting (reporting bias).

• Other bias.

We assessed outcome reporting bias (Kirkham 2010) by integrating the results of the table 'High risk of outcome reporting bias according to ORBIT classification' (Appendix 6) and the table 'Matrix of study endpoints (publication and trial documents'; Appendix 5). This analysis formed the basis for the judgement of selective reporting (reporting bias).

We judged risk of bias criteria as 'low risk', 'high risk' or 'unclear risk' and evaluated individual bias items as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a). We present a 'Risk of bias' graph and a 'Risk of bias summary' table .

We assessed the impact of individual bias domains on study results at endpoint and study levels.

For blinding of participants and personnel (performance bias), detection bias (blinding of outcome assessors) and attrition bias (incomplete outcome data) we intended to evaluate risk of bias separately for subjective and objective outcomes (Hróbjartsson 2013). We considered the implications of missing outcome data from individual participants.

We defined the following endpoints as subjective outcomes.

• Regression of Graves' ophthalmopathy.

• Health‐related quality of life.

We defined the following outcomes as objective outcomes.

• Rate of recurrent hyperthyroidism.

• Permanent recurrent laryngeal nerve palsy (if confirmed by laryngoscopy).

• Permanent hypocalcaemia/hypoparathyroidism.

• All‐cause mortality.

• Socioeconomic effects.

Measures of treatment effect

We expressed dichotomous data as odds ratios (ORs) or risk ratios (RRs) with 95% confidence intervals (CIs). We expressed continuous data as mean differences (MDs) with 95% CIs.

Unit of analysis issues

We took into account the level at which randomisation occurred, such as crossover trials, cluster‐randomised trials and multiple observations for the same outcome.

Dealing with missing data

We obtained relevant missing data from authors, if feasible, and evaluated important numerical data such as screened, eligible, randomised participants as well as intention‐to‐treat, as‐treated and per‐protocol populations. We investigated attrition rates, for example losses to follow‐up and withdrawals, and critically appraised issues of missing data and imputation methods (e.g. last observation carried forward).

Assessment of heterogeneity

In the event of substantial clinical, methodological or statistical heterogeneity, we did not report study results as meta‐analytically pooled effect estimates.

We identified heterogeneity by visual inspection of the forest plots and by using a standard Chi² test with a significance level of α = 0.1, in view of the low power of this test. We examined heterogeneity using the I² statistic, which quantifies inconsistency across studies to assess the impact of heterogeneity on the meta‐analysis (Higgins 2002; Higgins 2003), where an I² statistic of 75% or more indicates a considerable level of inconsistency (Higgins 2011a).

Had we found heterogeneity, we would have attempted to determine potential reasons for it by examining individual study and subgroup characteristics.

Assessment of reporting biases

In case of 10 studies or more for a given outcome, we planned to use funnel plots to assess small study effects. Due to several explanations for funnel plot asymmetry we planned to interpret results carefully (Sterne 2011).

Data synthesis

Unless there was good evidence for homogeneous effects across studies we planned to primarily summarise low risk of bias data by means of a random‐effects model (Wood 2008). We would have interpreted random‐effects meta‐analyses with due consideration to the whole distribution of effects, ideally by presenting a prediction interval (Higgins 2009). A prediction interval specifies a predicted range for the true treatment effect in an individual study (Riley 2011). In addition, we performed statistical analyses according to the statistical guidelines referenced in the latest version of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a).

Quality of evidence

We present the overall quality of the evidence for each outcome according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach, which takes into account issues related not only to internal validity (risk of bias, inconsistency, imprecision, publication bias) but also to external validity, such as directness of results (Atkins 2004). Two review authors (ZWL, LM) independently rated the quality of evidence for each outcome. We present a summary of the evidence in the 'Summary of findings' tables. These tables provide key information about the best estimate of the magnitude of the effect, in relative terms and as absolute differences, for each relevant comparison of alternative management strategies, numbers of participants and studies addressing each important outcome and rating of overall confidence in effect estimates for each outcome. We created the 'Summary of findings' tables on the basis of methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a) using RevMan's table editor. Otherwise, we would have used the GRADEproGDT software (GRADEproGDT 2015). We present results for the outcomes as described in the Types of outcome measures section. If meta‐analysis was not possible, we present the results in a narrative format in the 'Summary of findings' table. We justified all decisions to down‐ or up‐grade the quality of studies using footnotes, and we made comments to aid the reader's understanding of the review where necessary.

Subgroup analysis and investigation of heterogeneity

We planned to perform the following subgroup analyses and investigate interaction.

• Age (depending on data).

• Gender.

• Effects of previous treatment.

Sensitivity analysis

We planned to perform sensitivity analyses to explore the influence of the following factors (when applicable) on effect sizes by restricting the analysis to the following.

• Published trials.

• Taking into account risk of bias, as specified in the 'Assessment of risk of bias in included studies' section.

• Very long or large trials to establish the extent to which they dominate the results.

• Trials using the following filters: diagnostic criteria, imputation, language of publication, source of funding (industry versus other), or country.

We also planned to test the robustness of results by repeating the analysis using different measures of effect size (RR, OR, etc) and different statistical models (fixed‐effect and random‐effects models).

Results

Description of studies

Results of the search

We identified 694 records from our comprehensive literature searches; from these, we identified 46 full text papers for further examination. We excluded the other studies on the basis of their titles or abstracts because they did not meet the inclusion criteria or were not relevant to the question under study (see Figure 1 for the amended PRISMA (preferred reporting items for systematic reviews and meta‐analyses) flow diagram. After screening the full text of the selected publications, six trials (one of which is an ongoing study) met the inclusion criteria. All studies were published in English. We sought additional information from the lead author of the ongoing trial (Maschuw 2012) but unfortunately no preliminary data could be provided.

1.

Study flow diagram.

Included studies

A detailed description of the characteristics of included studies is presented elsewhere (see Characteristics of included studies and appendices). The following is a succinct overview.

Source of data

We obtained the data presented in the review from published literature and information in trial registers. We contacted authors for clarification and to obtain preliminary data from ongoing studies, however no new data were made available by correspondence.

Comparisons

Three thyroidectomy techniques were compared in the identified trials. Total thyroidectomy is removal of all macroscopic thyroid tissue. Two techniques of subtotal thyroidectomy are a) bilateral subtotal thyroidectomy, in which small tissue remnants (1 mL to 4 mL) are left on both sides of the thyroid gland, with the theoretical advantage of preventing postoperative hypothyroidism and permanent hypocalcaemia/hypoparathyroidism; and b) the Dunhill procedure, in which one lobe of the thyroid is entirely removed and a small remnant (1 mL to 4 mL) is left on the contralateral side.

Comparisons identified in the review fall into the following categories.

1. Total thyroidectomy versus bilateral subtotal thyroidectomy versus. Dunhill procedure (one RCT).

2. Total thyroidectomy versus bilateral subtotal thyroidectomy (one RCT).

3. Bilateral subtotal thyroidectomy versus Dunhill procedure (three RCTs).

Overview of study populations

A total of 886 participants were included in the five trials; 172 participants were randomised to total thyroidectomy, 383 were randomised to bilateral subtotal thyroidectomy, 309 were randomised to the Dunhill procedure and 22 were randomised to either bilateral subtotal thyroidectomy or the Dunhill procedure (not specified which procedure had been carried out).

A total of 165 (96%) participants completed the study in the total thyroidectomy group. Altogether 362 (94.5%) participants finished in the bilateral subtotal thyroidectomy group. A total of 274 (88.7%) completed follow‐up in the Dunhill procedure group and 22 (100%) completed follow‐up in the subtotal thyroidectomy group (surgical technique unspecified).

The individual sample size ranged from 44 to 340.

Study design

Studies were RCTs. Five trials adopted a parallel group superiority design and no studies contained a placebo group. There were no multicentre trials. One study had three comparison arms (Witte 2000).

In terms of blinding, there were no double‐blinded studies. One study was single‐blinded for participants (Barczynski 2012), and in four studies, blinding was not defined (Chi 2005; Jarhult 2005; Muller 2001; Witte 2000). Outcome assessors were blinded in two studies (Barczynski 2012; Jarhult 2005).

Studies were performed between the years 1988 to 2004. The duration of follow‐up ranged from three months to six years, with a mean follow‐up period of 40 months. No studies had a run‐in period. No studies were terminated prematurely.

Settings

All five studies were conducted in university hospitals or tertiary referral centres for thyroid disease. The countries involved were Germany, Sweden, Poland and Taiwan. None were performed in an outpatient setting.

Participants

The participating population consisted of the following: patients diagnosed with Graves' disease undergoing surgical treatment. All trials included participants from economically developed countries. There was no breakdown of ethnic group data. Most patients received a course of medical treatment for Graves' disease. This was reported in five trials (Barczynski 2012; Chi 2005; Jarhult 2005; Muller 2001; Witte 2000), with the mean duration of prior treatment being 9 to 34.2 months. Participants’ gender showed a 6:1 female predominance consistent with the population prevalence of Graves' disease. The mean age of participants in the trials ranged from 30 to 46 years. Comorbidities were not reported for the included studies. Five trials described cointerventions in participants in the form of preoperative anti‐thyroid medication to render the patients euthyroid (Barczynski 2012; Chi 2005; Jarhult 2005; Muller 2001; Witte 2000).

Criteria for entry into the individual trials are outlined in the Characteristics of included studies. Major exclusion criteria were 1) suspicion of thyroid malignancy, 2) thyroid remnant less than 2.5 x 1 x 1cm (Chi 2005) or previous total thyroidectomy (for studies that only compared subtotal thyroidectomy techniques), 3) preoperative recurrent laryngeal nerve palsy and 4) patient refusal.

Diagnosis

Participants were diagnosed with Grave's disease in all trials. All trials confirmed the diagnosis of Graves' disease on clinical and biochemical criteria, which is hyperthyroidism with endocrine ophthalmopathy and thyroid‐stimulating hormone receptor antibodies or thyrotropin binding inhibiting immunoglobulin antibodies. Two trials used an ultrasound scan in addition to the criteria above (Chi 2005; Muller 2001). Histological confirmation was used in two trials (Chi 2005; Jarhult 2005).

Interventions

All five trials reported treatment before the start of the trial which included anti‐thyroid medication (e.g. carbimazole) and beta‐blockers. All participants were rendered euthyroid prior to surgery. The duration of the intervention was the time taken to perform the specified type of thyroidectomy (knife to skin to surgical closure of wound). All trials used adequate interventions and comparators.

Outcomes

One study explicitly stated primary and secondary endpoints in the publication (Barczynski 2012), four studies reported endpoints without specifying whether they were primary or secondary. The most commonly defined primary outcomes in publications were recurrent hyperthyroidism and change in Graves' ophthalmopathy. Other commonly reported outcomes were temporary/permanent hypocalcaemia and temporary/permanent recurrent laryngeal nerve palsy.

Reporting of endpoints

A total of five trials collected a median of six (range six to eight) outcomes. Recurrent hyperthyroidism was measured in four trials (Barczynski 2012; Chi 2005; Muller 2001; Witte 2000). Grave's ophthalmopathy was assessed in three trials (Barczynski 2012; Jarhult 2005; Witte 2000). Postoperative autoimmune antibody levels were reported in two trials (Jarhult 2005; Witte 2000). Surgical complications such as postoperative hypoparathyroidism and recurrent laryngeal nerve palsy were measured in all trials; three trials also reported haematoma and wound complication rates (Chi 2005; Jarhult 2005; Muller 2001). One death was reported which occurred three years after surgery and was not reported to be related to the surgery or Graves' disease. No trial specifically reported health‐related quality of life or socioeconomic effects. For a summary of all outcomes assessed in each trial, see Appendix 5.

Excluded studies

We had to excluded a total of 38 trials (40 publications) after careful evaluation of the full publication (see Figure 1).

The main reasons for exclusion were lack of randomisation (n = 11), retrospective study (n = 12) and no direct comparison of two surgical techniques (n = 10). For further details, see Characteristics of excluded studies.

Risk of bias in included studies

For details of risk of bias in included trials see Characteristics of included studies. For an overview of review authors' judgements about each risk of bias item for individual studies and across all trials see Figure 2 and Figure 3. We investigated performance bias, detection bias and attrition bias separately for objective and subjective outcome measures. 'Objective outcome' measures were defined as rate of recurrent hyperthyroidism, permanent recurrent laryngeal nerve palsy (if confirmed by laryngoscopy), permanent hypocalcaemia/hypoparathyroidism, all‐cause mortality and socioeconomic effects.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

'Subjective outcome' measures were defined as regression of Graves' ophthalmopathy and health‐related quality of life.

Allocation

All five trials were described as randomised, although only one study (Barczynski 2012) specified that random numbers were computer generated and sealed envelopes were used for allocation concealment. We could not obtain any more information on randomisation methods for the remaining trials.

Blinding

Only one trial explicitly stated that blinding of the participants was undertaken (Barczynski 2012). The surgical team was not blinded in any of the studies. One trial reported blinding of outcome assessment for some outcome measures (Barczynski 2012). External, blinded personnel undertook thyroid function, calcium measurements and assessment of recurrent laryngeal nerve status. Two trials reported blinding of subjective outcome assessment (Barczynski 2012; Jarhult 2005). An external blinded ophthalmologist undertook assessment of Graves' ophthalmopathy. In one trial it was unclear whether any of the outcome assessments were blinded (Muller 2001). None of the trials provided sufficient information about blinding procedures.

Incomplete outcome data

Participant withdrawal figures were described in four trials that had losses to follow‐up (Barczynski 2012; Chi 2005; Jarhult 2005; Muller 2001). Analysis was reported as intention‐to‐treat in four trials (Barczynski 2012; Jarhult 2005; Muller 2001; Witte 2000). No intention‐to‐treat analysis was undertaken in the trial by Chi 2005. One trial did not report losses to follow‐up (Witte 2000). Detailed descriptions of participants' withdrawals and reasons underpinning them were not provided in any of the included trials. One trial had attrition rates with possible impact on the outcomes (Muller 2001, overall loss to follow‐up 35%).

Selective reporting

Risk of reporting bias was considered to be low in two trials (Barczynski 2012; Muller 2001). Where the trial measured primary outcomes of the review (recurrent hyperthyroidism and adverse events), the results were reported in all but one trial (Jarhult 2005) as thyroid function was measured but not reported. Three trials did not explicitly report all‐cause mortality (Chi 2005; Jarhult 2005; Witte 2000).

Other potential sources of bias

We did not identify any other potential sources of bias.

Effects of interventions

See: Table 1; Table 2; Table 3

Baseline characteristics

For details of baseline characteristics, see Appendix 3 and Appendix 4.

1. Total thyroidectomy compared to subtotal thyroidectomy (both bilateral subtotal and the Dunhill procedure)

For recurrent hyperthyroidism and regression of Graves' ophthalmopathy, the results are based on a meta‐analysis of two studies (Barczynski 2012; Witte 2000). For hypoparathyroidism and recurrent laryngeal nerve palsy, the results are based on a meta‐analysis of three studies (Barczynski 2012; Jarhult 2005; Witte 2000).

Primary outcomes

Rate of recurrent hyperthyroidism

Total thyroidectomy was more effective than subtotal thyroidectomy at preventing recurrent hyperthyroidism: OR 0.14 (95% confidence interval (CI) 0.04 to 0.46); P = 0.001; 350 participants; 2 trials; moderate quality evidence (Figure 4). These findings translate into 8 (95% CI 2 to 26) per 1000 versus 55 per 1000 individuals with recurrent hyperthyroidism, if treated by total versus subtotal thyroidectomy, respectively.

4.

Forest plot of comparison: 2 Total thyroidectomy versus subtotal thyroidectomy (both bilateral subtotal thyroidectomy and the Dunhill procedure), outcome: 2.1 Recurrent hyperthyroidism.

TT: total thyroidectomy, ST: subtotal thyroidectomy

Adverse events: permanent hypocalcaemia/hypoparathyroidism, permanent recurrent laryngeal nerve palsy and wound complications

Total thyroidectomy conferred a greater risk of permanent hypocalcaemia/hypoparathyroidism: OR 4.79 (95% CI 1.36 to 16.83); P = 0.01; 393 participants; 3 trials; low quality evidence (Analysis 1.2). These findings translate into 59 (95% CI 16 to 190) per 1000 versus 14 per 1000 individuals with permanent hypocalcaemia/hypoparathyroidism, if treated by total versus subtotal thyroidectomy, respectively.

1.2. Analysis.

Comparison 1 Total thyroidectomy versus subtotal thyroidectomy, Outcome 2 Permanent hypocalcaemia/hypoparathyroidism.

There were no marked differences in rates of recurrent laryngeal nerve palsy between the total and subtotal thyroidectomy groups: OR 1.45 (95% CI 0.38 to 5.59); P = 0.59; 393 participants; 3 trials; low quality evidence (Analysis 1.3).

1.3. Analysis.

Comparison 1 Total thyroidectomy versus subtotal thyroidectomy, Outcome 3 Permanent recurrent laryngeal nerve palsy.

One trial reported a postoperative haematoma rate of 4.5% in the total thyroidectomy group versus 9.5% in the subtotal thyroidectomy group (Jarhult 2005). However, the absolute numbers in this trial were small.

Regression of Graves' ophthalmopathy

There were no marked differences in postoperative regression of Graves' ophthalmopathy between the total and subtotal thyroidectomy groups: OR 1.15 (95% CI 0.64 to 2.08); P = 0.64; 229 participants; 2 trials; low quality evidence (Analysis 1.4).

1.4. Analysis.

Comparison 1 Total thyroidectomy versus subtotal thyroidectomy, Outcome 4 Regression of Graves' ophthalmopathy.

All‐cause mortality

Only a single death was reported in the total thyroidectomy group which was seen as unrelated to the surgical intervention (myocardial infarction three years after surgery) (Barczynski 2012).

Socioeconomic effects

None of the trials reported information relevant to these outcomes.

Health‐related quality of life

None of the trials reported information relevant to health‐related quality of life.

2. Total thyroidectomy compared to bilateral subtotal thyroidectomy

This is based on a meta‐analysis of two trials with a total of 300 participants (Barczynski 2012; Witte 2000).

Primary outcomes

Rate of recurrent hyperthyroidism

Total thyroidectomy was more effective than bilateral subtotal thyroidectomy at preventing recurrent hyperthyroidism: OR 0.13 (95% CI 0.04 to 0.44); P = 0.001; 300 participants; 2 trials; moderate quality evidence (Figure 5). These findings translate into 9 (95% CI 3 to 30) per 1000 versus 67 per 1000 individuals with recurrent hyperthyroidism, if treated by total versus bilateral subtotal thyroidectomy, respectively.

5.

Forest plot of comparison: 1 Total thyroidectomy versus bilateral subtotal thyroidectomy, outcome: 1.1 Recurrent hyperthyroidism.

TT: total thyroidectomy, BST: bilateral subtotal thyroidectomy

Adverse events: permanent hypocalcaemia/hypoparathyroidism and permanent recurrent laryngeal nerve palsy

No marked differences were seen in rates of permanent hypocalcaemia/hypoparathyroidism: OR 2.04 (95% CI 0.53 to 7.85); P = 0.30; 300 participants; 2 trials low quality evidence (Analysis 2.2) or permanent recurrent laryngeal nerve palsy between the comparator groups: OR 1.00 (95% CI 0.14 to 7.15); P = 1.0; 300 participants; 2 trials; low quality evidence (Analysis 2.3).

2.2. Analysis.

Comparison 2 Total thyroidectomy versus bilateral subtotal thyroidectomy, Outcome 2 Permanent hypocalcaemia/hypoparathyroidism.

2.3. Analysis.

Comparison 2 Total thyroidectomy versus bilateral subtotal thyroidectomy, Outcome 3 Permanent recurrent laryngeal nerve palsy.

Secondary outcomes

Regression of Graves' ophthalmopathy

There were no marked differences in postoperative regression of Graves' ophthalmopathy between the total and bilateral subtotal thyroidectomy groups: OR 1.19 (95% CI 0.64 to 2.21); P = 0.58; 260 participants; 2 trials; low quality evidence (Analysis 2.4).

2.4. Analysis.

Comparison 2 Total thyroidectomy versus bilateral subtotal thyroidectomy, Outcome 4 Regression of Graves' ophthalmopathy.

All‐cause mortality

Only a single death was reported in the total thyroidectomy group (Barczynski 2012) which was seen as unrelated to the surgical intervention (myocardial infarction three years after surgery).

Socioeconomic effects

None of the trials reported information relevant to these outcomes.

Health‐related quality of life

None of the trials reported information relevant to health‐related quality of life.

3. Bilateral subtotal thyroidectomy compared to the Dunhill procedure

Three trials provided data on outcome measures (Chi 2005; Muller 2001; Witte 2000).

Primary outcomes

Rate of recurrent hyperthyroidism

Patients in the bilateral subtotal thyroidectomy groups were more likely than those in the Dunhill procedure group to experience recurrent hyperthyroidism: OR 2.73 (95% CI 1.28 to 5.85); P = 0.01; 592 participants; 3 trials; low quality evidence (Figure 6). These findings translate into 68 (95% CI 33 to 135) per 1000 versus 26 per 1000 individuals with recurrent hyperthyroidism, if treated by bilateral subtotal thyroidectomy versus the Dunhill procedure, respectively.

6.

Forest plot of comparison: 3 Bilateral subtotal thyroidectomy versus the Dunhill procedure, outcome: 3.1 Recurrent hyperthyroidism.

BST: bilateral subtotal thyroidectomy

Adverse events: permanent hypocalcaemia/hypoparathyroidism, permanent recurrent laryngeal nerve palsy and wound complications

There were no marked differences between the bilateral subtotal thyroidectomy and the Dunhill procedure groups in rates of permanent hypocalcaemia/hypoparathyroidism: OR 1.90 (95% CI 0.47 to 7.71); P = 0.37; 592 participants; 3 trials; low quality evidence (Analysis 3.2).

3.2. Analysis.

Comparison 3 Bilateral subtotal thyroidectomy versus the Dunhill procedure, Outcome 2 Permanent hypocalcaemia/hypoparathyroidism.

There were no marked differences between the bilateral subtotal thyroidectomy and the Dunhill procedure groups in rates of permanent recurrent laryngeal nerve palsy: (OR 0.14 (95% CI 0.00 to 6.82); P = 0.32; 592 participants; 3 trials; low quality evidence (Analysis 3.3).

3.3. Analysis.

Comparison 3 Bilateral subtotal thyroidectomy versus the Dunhill procedure, Outcome 3 Permanent recurrent laryngeal nerve palsy.

Overall rates of haematoma and other wound complications related to surgery were low. Postoperative haematoma requiring drainage or return to the theatre were in the range of 0.5% to 3% (Chi 2005), and seroma rates were 3% to 4.7% (Muller 2001). Direct comparison between the studies was not possible due to different reporting criteria.

Secondary outcomes

Regression of Graves' ophthalmopathy

There were no marked differences between the bilateral subtotal thyroidectomy and the Dunhill procedure groups in regression of Graves' ophthalmopathy postoperatively: OR 0.78 (95% CI 0.32 to 1.92); P = 0.59; 90 participants; 2 trials; low quality evidence (Analysis 3.4).

3.4. Analysis.

Comparison 3 Bilateral subtotal thyroidectomy versus the Dunhill procedure, Outcome 4 Regression of Graves' ophthalmopathy.

All‐cause mortality

No cases of mortality were reported in any of the relevant trials.

Socioeconomic effects

None of the trials reported information relevant to these outcomes.

Health‐related quality of life

None of the trials reported information relevant to health‐related quality of life.

Subgroup analyses

We did not perform subgroups analyses because there were not enough trials to estimate effects in various subgroups.

Sensitivity analyses

We did not perform any sensitivity analyses because there were not enough trials.

Assessment of reporting bias

We did not draw funnel plots due to the limited number of trials (n = 5).

Discussion

Summary of main results

Total thyroidectomy was more effective than subtotal thyroidectomy techniques (both bilateral subtotal thyroidectomy and the Dunhill procedure (unilateral total and contralateral subtotal thyroidectomy)) at preventing recurrent hyperthyroidism (moderate quality evidence). There were neutral effects for regression of Graves' ophthalmopathy and permanent recurrent laryngeal nerve palsy when comparing total with subtotal thyroidectomy, but increased rates of hypocalcaemia/hypoparathyroidism when comparing total with subtotal thyroidectomy (all low quality evidence).

The Dunhill procedure compared with bilateral subtotal thyroidectomy was more effective in preventing recurrent hyperthyroidism (low quality evidence). There were neutral effects for regression of Graves' opthalmopathy, permanent recurrent laryngeal nerve palsy and permanent hypocalcaemia/hypoparathyroidism when comparing these two techniques (all low quality evidence).

Deaths were rarely reported, health‐related quality of life and socioeconomic effects were not investigated in any of the included studies.

Overall completeness and applicability of evidence

We identified five studies and one ongoing study in our comprehensive literature search that fit the criteria for inclusion. We carried out a meta‐analysis to compare outcomes from three types of interventions We did not exclude any studies based on language of publication. We sought data from the authors where study results were incompletely reported.

The main conclusions from the meta‐analysis can be extrapolated to general clinical practice, however it is noted that all studies were carried out in institutions with high case throughput in thyroidectomies. It is possible that complication rates (such as permanent recurrent laryngeal nerve palsy or permanent hypocalcaemia/hypoparathyroidism) may be higher where surgeons do not perform thyroidectomy for Graves' disease on a regular basis.

Quality of the evidence

Overall quality of the evidence ranged from low to moderate. All trials were randomised. However, in all the trials blinding procedures were insufficiently described. Objective outcome assessment was blinded in two trials (Barczynski 2012; Jarhult 2005). Surgeons were not blinded in any of the trials. One trial blinded participants (Barczynski 2012). Attrition bias applied to one trial (Muller 2001), with 35% loss to follow‐up. In one trial the analysis was not intention‐to‐treat (Chi 2005).

Quality of evidence was highest when surgical techniques were compared in terms of postoperative recurrent hyperthyroidism, as the outcome is objective. Quality of evidence was lowest when rates of complications (recurrent laryngeal nerve palsy and permanent hypocalcaemia/hypoparathyroidism) were considered (as event rates were low across all studies). Overall, the main reasons for downgrading decisions in the 'Summary of findings' tables were high risk of performance bias, imprecision and low event rates.

Potential biases in the review process

The search for this topic was comprehensive and fully up‐to‐date, allowing for identification of all appropriate studies. We were unable to obtain additional information from authors on methodology; therefore risk of bias is unclear in some studies. We were only able to compare one study protocol with its publication (Barczynski 2012), therefore evaluation of reporting bias was limited.

Agreements and disagreements with other studies or reviews

Our results are consistent with the systematic review and meta‐analysis conducted by Palit et al, who found an 8% rate of recurrent hyperthyroidism with subtotal thyroidectomy (compared with total thyroidectomy), with no marked differences in complication rates (Palit 2000). A systematic review by Guo et. al. came to similar conclusions indicating that total thyroidectomy was more effective than subtotal thyroidectomy at preventing recurrent hyperthyroidism, although total thyroidectomy was associated with a higher rates of temporary hypoparathyroidism (Guo 2013). No marked differences were found in rates of permanent hypocalcaemia/hypoparathyroidism or recurrent laryngeal nerve palsy.

Authors' conclusions

Implications for practice.

Total thyroidectomy is more effective than subtotal thyroidectomy (both bilateral subtotal thyroidectomy and the Dunhill procedure (unilateral total and contralateral subtotal thyroidectomy)) at preventing recurrent hyperthyroidism in Graves' disease. Complications like permanent recurrent nerve palsy or permanent hypocalcaemia/hypoparathyroidism were not clearly associated with one of these techniques. However, there was some evidence for increased rates of permanent hypocalcaemia/hypoparathyroidism when comparing total with subtotal thyroidectomy. Effects were neutral for regression of Graves' ophthalmopathy. The Dunhill procedure compared with bilateral subtotal thyroidectomy was more effective in preventing recurrent hyperthyroidism, effects were neutral for complications and regression of Graves' ophthalmopathy. Deaths were rarely reported, health‐related quality of life and socioeconomic effects were not investigated in any of the included studies.

Implications for research.

There was a lack of high quality evidence on postoperative complications of thyroidectomy to conclude whether total thyroidectomy carries a higher risk of permanent hypocalcaemia/hypoparathyroidism or recurrent laryngeal nerve palsy compared to subtotal thyroidectomy techniques. In our included studies, there were no measures of health‐related quality of life or socioeconomic effects associated with thyroid surgery for Graves' disease. Well designed trials that are randomised, adequately powered and have suitable follow‐up are required to assess the benefit‐risk ratio of treatments for control of Graves hyperthyroidism and Graves' opthalmopathy. Of particular interest are adverse events and morbidity; standardised reporting will be important in this respect. Blinding of participants and personnel as well as outcome assessors for most outcome measures appears essential and all trials should be prospectively registered to analyse reporting bias amongst other things. Finally, health‐related quality of life is an important outcome and should form a core determinant of any future study.

Notes

We have based parts of the background, the methods section, appendices, additional tables and figures 1 to 3 of this review on a standard template established by the Cochrane Metabolic and Endocrine Disorders (CMED) Group.

Appendices

Appendix 1. Search strategies

Cochrane Library

1. [mh "Graves disease"] 2. (grave* near/7 (diseas* or thyrotoxicos* or hyperthyr*)):ti,ab,kw 3. (grave* near/7 (orbitopath* or ophthalmopath*)):ti,ab,kw 4. (basedow* near/7 (diseas* or syndrom*)):ti,ab,kw 5. {or #1‐#4} 6. [mh Thyroidectomy] 7. thyroidectom*:ti,ab,kw 8. #6 or #7 9. #5 and #8

MEDLINE (Ovid SP)

1. exp Graves Disease/ 2. (grave* adj6 (diseas* or thyrotoxicos* or hyperthyr*)).tw. 3. (grave* adj6 (orbitopath* or ophthalmopath*)).tw. 4. (basedow* adj6 (diseas* or syndrom*)).tw. 5. or/1‐4 6. exp Thyroidectomy/ 7. thyroidectom*.tw. 8. 6 or 7 9. 5 and 8 [10‐19:Lefebvre 2011RCT filter ‐ max. sensitivity version, without "drug therapy.fs."] 10. randomized controlled trial.pt. 11. controlled clinical trial.pt. 12. randomized.ab. 13. placebo.ab. 14. randomly.ab. 15. trial.ab. 16. groups.ab. 17. or/ 10‐16 18. exp animals/ not humans/ 19. 17 not 18 20. 9 and 19

PubMed (subsets not available on Ovid)

1. grave*[tw] AND (diseas*[tw] OR thyrotoxicos*[tw] OR hyperthyr*[tw] OR orbitopath*[tw] OR ophthalmopath*[tw]) 2. basedow*[tw] AND (diseas*[tw] OR syndrom*[tw]) 3. #1 OR #2 4. thyroidectom*[tw] OR surg*[tw] 5. #3 AND #4 6. random*[tw] OR trial[tw] OR groups[tw] 7. #5 AND #6 8. #7 NOT medline[sb] NOT pmcbook

EMBASE (Ovid SP)

1. exp Graves disease/ 2. (grave* adj6 (diseas* or thyrotoxicos* or hyperthyr*)).tw. 3. (grave* adj6 (orbitopath* or ophthalmopath*)).tw. 4. (basedow* adj6 (diseas* or syndrom*)).tw. 5. or/1‐ 4 6. exp subtotal thyroidectomy/ or exp thyroidectomy/ 7. thyroidectom*.tw. 8. 6 or 7 9. 5 and 8 [10:Wong 2006"treatment studies" filter ‐ SDSSGS version] 10. random*.tw. or clinical trial*.mp. or exp treatment outcome/ 11. 9 and 10 12. limit 11 to embase

ICTRP Search Portal (Standard search)

grave* AND thyroidectom* OR basedow* AND thyroidectom* OR grave* AND surg* OR basedow* AND surg*

ClinicalTrials.gov (Advanced search)

Conditions: grave OR graves OR basedow OR basedows OR "exophthalmic goiter" OR "exophthalmic goiters" OR "autoimmune hyperthyroidism" Interventions: thyroidectomy OR thyroidectomies OR surgery OR surgical OR surgeries

Appendix 2. Description of interventions

	Intervention	Comparator intervention
Barczynski 2012	Total thyroidectomy	Bilateral subtotal thyroidectomy
Witte 2000	Total thyroidectomy	Bilateral subtotal thyroidectomy
		Unilateral total and contralateral subtotal thyroidectomy (Dunhill procedure)
Jarhult 2005	Total thyroidectomy	Subtotal thyroidectomy
Chi 2005	Bilateral subtotal thyroidectomy	Unilateral total and contralateral subtotal thyroidectomy (Dunhill procedure)
Muller 2001	Bilateral subtotal thyroidectomy	Unilateral total and contralateral subtotal thyroidectomy (Dunhill procedure)

Appendix 3. Baseline characteristics (I)

	Intervention and comparator intervention	Duration of follow‐up	Participating population	Study period [year to year]	Country	Setting	Ethnic groups [%]	Duration of Graves' disease
Barczynski 2012	I: total thyroidectomy	5 years	Adults diagnosed with Graves' disease and mild active ophthalmopathy	2000 to 2004	Poland	Teaching hospital	‐	< 24 months
	C: bilateral subtotal thyroidectomy							< 24 months
Witte 2000	I: total thyroidectomy	6 months	Participants with clinically or biochemically proven Graves' disease	1993 to 1995	Germany	Teaching hospital	‐	2‐300 months
	C1: bilateral subtotal thyroidectomy							2‐300 months
	C2: Dunhill procedure							2‐300 months
Jarhult 2005	I: total thyroidectomy	3 years	Participants with Graves' disease and moderate to severe ophthalmopathy	‐	Sweden	University/county hospitals	‐	‐
	C: subtotal thyroidectomy
Chi 2005	I: bilateral subtotal thyroidectomy	26.4 (SD 1.1) months	Participants undergoing subtotal thyroidectomy for Graves' hyperthyroidism	1998 to 2002	Taiwan	Teaching hospital	‐	‐
	C: Dunhill procedure
Muller 2001	I: bilateral subtotal thyroidectomy	6 years	Diagnosed with Graves' disease if participants had hyperthyroidism and had ophthalmopathy or TSH‐R antibodies, or based on US scan	1988 to 1992	Germany	Tertiary referral centre	‐	‐
	C: Dunhill procedure
‐ denotes not reported C: comparator; I: intervention; SD: standard deviation; TSH‐R: thyroid‐stimulating hormone receptor; US: ultrasound

Appendix 4. Baseline characteristics (II)

	Intervention and comparator intervention	Sex [female %]	Age [mean years (CI/SD/range)]	Comedications/Cointerventions	Comorbidities
Barczynski 2012	I: total thyroidectomy	89	46.2 (CI 43.5 to 48.9)	Antithyroid drugs/beta blockers	‐
	C: bilateral subtotal thyroidectomy	89	45.5 (CI 43.1 to 47.9)
Witte 2000	I: total thyroidectomy	82	38 (19 to 68)	Antithyroid drugs/beta blockers (40%), thyroid hormone substitution (40.7%)	‐
	C1: bilateral subtotal thyroidectomy	86	35.5 (12 to 73)
	C2: Dunhill procedure	82	41 (17 to 66)
Jarhult 2005	I: total thyroidectomy	95	44 (21 to 62)	Antithyroid drugs for at least 3 months preoperatively Postoperative thyroxine supplementation 5 participants in each group received cortisone to decrease the activity of the EO	‐
	C: subtotal thyroidectomy	86	42 (24 to 69)
Chi 2005	I: bilateral subtotal thyroidectomy	84	33.3 (SD 0.73)	Propylthiouracil/methimazole	‐
	C: Dunhill procedure	87	34.5 (SD 0.93)
Muller 2001	I: bilateral subtotal thyroidectomy	‐	30.8 (14 to 78)	Antithyroid medication	‐
	C: Dunhill procedure	‐	30.8 (14 to 78)
‐ denotes not reported C: comparator; CI: confidence interval; EO: endocrine ophthalmopathy; I: intervention; SD: standard deviation

Appendix 5. Matrix of study endpoints (publications and trial documents)

	Endpoints quoted in trial document(s) (ClinicalTrials.gov, FDA/EMA document, manufacturer's website, published design paper)a	Study results posted in trial register [Yes/No]	Publications specified in trial register	Endpoints quoted in publication(s)b,c	Endpoints quoted in abstract of publication(s)b,c
Barczynski 2012	Source: NCT01408368 Primary outcome measure(s): long term control of Graves' disease (time frame: up to 60 months postoperatively), recurrence rate of hyperthyroidism and change in Graves' ophthalmopathy	No	Yes	Primary outcome measure(s): prevalence of recurrent hyperthyroidism during follow‐up and changes in Graves’ ophthalmopathy	Primary outcome measure(s): prevalence of recurrent hyperthyroidism and changes in Graves’ ophthalmopathy
	Secondary outcome measure(s): morbidity rate (time frame: up to 12 months postoperatively), RLN injury and hypoparathyroidism			Secondary outcome measure(s): postoperative hypoparathyroidism and recurrent RLN injury	Secondary outcome measure(s): postoperative transient and permanent paresis of the RLN, postoperative hypocalcaemia and hypoparathyroidism
	Other outcome measure(s): ‐			Other outcome measure(s): serum free triiodothyronine, free thyroxine and thyroid‐stimulating hormone, serum thyrotropin binding inhibitory immunoglobulin (TBII) levels, ophthalmological assessment, clinical activity score to evaluate the activity of the ophthalmopathy, high‐resolution Doppler ultrasonography of the neck, duration of operation, parathyroid autotransplantation, parathyroid glands identified by histology in thyroid specimens, incidental papillary thyroid cancer, diameter of thyroid cancer	Other outcome measure(s):
Chi 2005	N/T			Primary outcome measure(s): ‐	Primary outcome measure(s): ‐
				Secondary outcome measure(s): ‐	Secondary outcome measure(s): ‐
				Other outcome measure(s): operative time, triiodothyronine (T3), thyroxine (T4), thyrotropin (TSH), and antimicrosomal antibodies, postoperative serum calcium, permanent hypoparathyroidism, complications (postoperative mortality, RLN palsy, wound haematoma or bleeding after thyroidectomy, need of calcium medication), recurrent hyperthyroidism, hypothyroidism, incidental carcinoma	Other outcome measure(s): operative time, recurrent hyperthyroidism, hypothyroidism, postoperative complications
Witte 2000	N/T			Primary outcome measure(s): ‐	Primary outcome measure(s): ‐
				Secondary outcome measure(s): ‐	Secondary outcome measure(s): ‐
				Other outcome measure(s): thyroid‐stimulating hormone receptor antibodies (TSH‐RAb), postoperative total remnant size measured by ultrasonography, clinical courses of Graves’ ophthalmopathy using the modified scale system of the American Thyroid Association (ATA), early postoperative complications (RLN palsy, hypothyroidism, wound infection) until discharge from hospital, long‐term results of changes in endocrine orbitopathy, serum calcium levels, vocal cord motility, and TSH‐RAb titers were measured between 18 and 58 months after thyroid operation	Other outcome measure(s): postoperative Graves’ ophthalmopathy changes, postoperative TSH‐RAb titers, postoperative complication rates
Jarhult 2005	N/T			Primary outcome measure(s): ‐	Primary outcome measure(s): ‐
				Secondary outcome measure(s): ‐	Secondary outcome measure(s): ‐
				Other outcome measure(s): evaluation of ophthalmopathy, patients overall opinion about the actual eye problems was expressed on a visual analogue scale, thyroid remnant, operating time, reoperation due to bleeding, hospital stay, permanent vocal cord palsy, permanent hypothyroidism, TSH, free triiodothyronine (T3) and free thyroxine (T4), thyrotropin receptor antibodies (TRAB) titre, calcium	Other outcome measure(s): development of ophthalmopathy, thyrotropin (TSH)‐receptor antibody levels, surgical complication rate (permanent RLN paresis and permanent hypoparathyroidism)
Muller 2001	N/T			Primary outcome measure(s): ‐	Primary outcome measure(s): ‐
				Secondary outcome measure(s): ‐	Secondary outcome measure(s): ‐
				Other outcome measure(s): recurrent hyperthyroidism, regression of Graves' ophthalmopathy, all‐cause mortality, adverse events	Other outcome measure(s): mortality, recurrent laryngeal nerve palsy, hypocalcaemia, secondary haemorrhage, wound infections, recurrence rate
‐ denotes not reported aTrial document(s) refers to all available information from published design papers and sources other than regular publications (e.g. FDA/EMA documents, manufacturer's websites, trial registers) bPublication(s) refers to trial information published in scientific journals (primary reference, duplicate publications, companion documents or multiple reports of a primary study) cOther outcome measures refer to all outcomes not specified as primary or secondary outcome measures EMA: European Medicines Agency; FDA: Food and Drug Administration (US); N/T: no trial document available; RLN: recurrent laryngeal nerve

Appendix 6. High risk of outcome reporting bias according to ORBIT classification

	Outcome	High risk of bias (category A)a	High risk of bias (category D)b	High risk of bias (category E)c	High risk of bias (category G)d
Barczynski 2012	N/A
Chi 2005	All‐cause mortality				x
Jarhult 2005	All‐cause mortality				x
	Rate of recurrent hyperthyroidism		x
Witte 2000	All‐cause mortality				x
Muller 2001	N/A
aClear that outcome was measured and analysed; trial report states that outcome was analysed but only reports that result was not significant (Classification 'A', table 2, Kirkham 2010). bClear that outcome was measured and analysed; trial report states that outcome was analysed but no results reported (Classification 'D', table 2, Kirkham 2010). cClear that outcome was measured but not necessarily analysed; judgement says likely to have been analysed but not reported because of non‐significant results (Classification 'E', table 2, Kirkham 2010). dUnclear whether the outcome was measured; not mentioned but clinical judgement says likely to have been measured and analysed but not reported on the basis of non‐significant results (Classification 'G', table 2, Kirkham 2010). N/A: not applicable; ORBIT: Outcome Reporting Bias In Trials.

Appendix 7. Definition of endpoint measurement

	Rate of recurrent hyperthyroidism	Permanent recurrent laryngeal nerve palsy	Permanent hypocalcaemia/hypoparathyroidism	Regression of Graves' ophthalmopathy	Health‐related quality of life	Socioeconomic effects	Severe/serious adverse events
Barczynski 2012	Laboratory tests included measurement of serum free tri‐iodothyronine (fT3; reference range 0.22 to 6.78 pmol/L), free thyroxine (fT4; reference range 12 to 30 pmol/L) and thyroid‐stimulating hormone (TSH, thyrotropin; reference range 0.4 to 4.2 units/L). A serum thyrotropin binding inhibitory immunoglobulin (TBII) level > 1.5 units/L was defined as positive	Vocal cord paresis for more than 12 months after the operation was regarded as permanent Indirect laryngoscopy was performed before surgery	Persistent hypocalcaemia with serum PTH level below 10 ng/L for more than 12 months	The following definitions were used: a change in TES and/or CAS of fewer than 2 points from the preoperative baseline value was regarded as ‘no change’, an increase in TES and/or CAS of 2 points or more was regarded as‘progression’, and a decrease in TES and/or CAS of 2 points or more was regarded as ‘improvement’ Assessed by blinded ophthalmologist Measured at yearly intervals for 5 years postop	N/I	N/I	One death reported (due to a myocardial infarction 3 years after surgery)
Chi 2005	Suppressed TSH (human TSH radioimmunoassay kit measured between 3‐26.4 months after surgery	RLN palsy defined as permanent if longer than 6 months Based on independent ENT specialist assessment and patient report	Permanent hypocalcaemia if longer than 6 months postop.. Assessed by external laboratory	N/I	N/I	N/I	N/I
Jarhult 2005	N/I	RLN palsy considered permanent if persisting > 9 months (unclear whether laryngoscopy was performed)	Persistent hypocalcaemia > 9 months	1. Improvement in ophthalmological score (based on author experience, not validated) 2. Improvment in objective and subjective evaluation of eye symptoms by patient and ophthalmologist	N/I	N/I	N/I
Witte 2000	N/I	RLN palsy defined as permanent if longer than 6 months Assessed by independent ENT specialist	Permanent hypocalcaemia if longer than 6 months postop.. Assessed by external laboratory	GO measured from > 18‐58 months postop using standardised scale Both patient reported and ophthalmology assessment scores are reported	N/I	N/I	N/I
Muller 2001	Biochemical tests of TSH, fT3 and fT4 with reference levels as per local laboratory. Suppressed TSH and elevated fT3/fT4 was taken to be evidence of hyperthyroidism (measured at 6 years)	Recurrent laryngeal nerve palsy defined by impaired vocal cord movement seen on laryngoscopy (preop and postop laryngoscopy routine). Considered permanent if persisting more than 1 year	Hypocalcaemia persisting more than 1 year was considered to be permanent	Surgical team/physician assessment Unclear what scale was used or whether an ophthalmologist was involved Outcome measured at 6 years	N/I	N/I	No deaths
CAS: clinical activity score; ENT: ear, nose & throat; fT3: free tri‐iodothyronine; fT4: free thyroxine; GO: Graves' ophthalmopathy; N/I: not investigated; PTH: parathyroid hormone; RLN: recurrent laryngeal nerve; TES: total eye score; TT: total thyroidectomy; TSH: thyroid‐stimulating hormone

Appendix 8. Adverse events (I)

	Intervention and comparator intervention	Randomised/safety [N]	Deaths [N]	Deaths [%]	All adverse events [N]	All adverse events [%]	Severe/serious adverse events [N]	Severe/serious adverse events [%]
Barczynski 2012	I: total thyroidectomy	100	1	1	21	21	2	2
	C: bilateral subtotal thyroidectomy	100	0	0	35	35	3	3
Witte 2000	I: total thyroidectomy	50	‐	‐	10	20	1	2
	C1: bilateral subtotal thyroidectomy	50	‐	‐	16	32	0	0
	C2: Dunhill procedure	50	‐	‐	9	18	1	2
Jarhult 2005	I: total thyroidectomy	22	‐	‐	5	23	‐	‐
	C: subtotal thyroidectomy	21	‐	‐	1	5	‐	‐
Chi 2005	I: bilateral subtotal thyroidectomy	166	‐	‐	22	13.3	0	0
	C: Dunhill procedure	174	‐	‐	55	31.6	1	0.6
Muller 2001	I: bilateral subtotal thyroidectomy	67	0	0	14	21	1	1.5
	C: Dunhill procedure	85	0	0	18	21	2	2.3
‐ denotes not reported C: comparator intervention; I: intervention

Appendix 9. Adverse events (II)

	Intervention and comparator intervention	Randomised/safety [N]	Left study due to adverse events [N]	Left study due to adverse events [%]	Hospitalisation [N]	Hospitalisation [%]	Outpatient treatment [N]	Outpatient treatment [%]
Barczynski 2012	I: total thyroidectomy	100	1	1	‐	‐	‐	‐
	C: bilateral subtotal thyroidectomy	100	0	0	‐	‐	‐	‐
Witte 2000	I: total thyroidectomy	50	‐	‐	‐	‐	‐	‐
	C1: bilateral subtotal thyroidectomy	50	‐	‐	‐	‐	‐	‐
	C2: Dunhill procedure	50	‐	‐	‐	‐	‐	‐
Jarhult 2005	I: total thyroidectomy	22	‐	‐	‐	‐	‐	‐
	C: subtotal thyroidectomy	21	‐	‐	‐	‐	‐	‐
Chi 2005	I: bilateral subtotal thyroidectomy	166	‐	‐	9	5.4	‐	‐
	C: Dunhill procedure	174	‐	‐	15	8.6	‐	‐
Muller 2001	I: bilateral subtotal thyroidectomy	67	‐	‐	‐	‐	‐	‐
	C: Dunhill procedure	85	‐	‐	‐	‐	‐	‐
‐ denotes not reported C: comparator; I: intervention

Appendix 10. Adverse events (III)

	Intervention and comparator intervention	Randomised/safety [N]	Specific adverse events [description]	Specific adverse events [N]	Specific adverse events [%]
Barczynski 2012	I: total thyroidectomy	100	‐	‐	‐
	C: bilateral subtotal thyroidectomy	100	‐	‐	‐
Witte 2000	I: total thyroidectomy	50	‐	‐	‐
	C1: bilateral subtotal thyroidectomy	50	‐	‐	‐
	C2: Dunhill procedure	50	‐	‐	‐
Jarhult 2005	I: total thyroidectomy	22	‐	‐	‐
	C: subtotal thyroidectomy	21	‐	‐	‐
Chi 2005	I: bilateral subtotal thyroidectomy	166	‐	‐	‐
	C: Dunhill procedure	174	‐	‐	‐
Muller 2001	I: bilateral subtotal thyroidectomy	67	‐	‐	‐
	C: Dunhill procedure	85	‐	‐	‐
‐ denotes not reported C: comparator; I: intervention

Appendix 11. Survey of study authors providing information on included trials

	Study author contacted	Study author replied	Study author asked for additional information	Study author provided data
Barczynski 2012	04/04/2014	05/04/2014	Yes	Yes
Witte 2000	04/04/2014	Did not reply	Yes	No
Jarhult 2005	24/11/2012	25/11/2012	Yes	Yes
Chi 2005	04/04/2014	Did not reply	Yes	No
Muller 2001	04/04/2014	Did not reply	Yes	No

Data and analyses

Comparison 1. Total thyroidectomy versus subtotal thyroidectomy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Rate of recurrent hyperthyroidism	2	350	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.14 [0.04, 0.46]
2 Permanent hypocalcaemia/hypoparathyroidism	3	393	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.79 [1.36, 16.83]
3 Permanent recurrent laryngeal nerve palsy	3	393	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.45 [0.38, 5.59]
4 Regression of Graves' ophthalmopathy	2	287	Odds Ratio (M‐H, Fixed, 95% CI)	1.15 [0.64, 2.08]

1.1. Analysis.

Comparison 1 Total thyroidectomy versus subtotal thyroidectomy, Outcome 1 Rate of recurrent hyperthyroidism.

Comparison 2. Total thyroidectomy versus bilateral subtotal thyroidectomy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Rate of recurrent hyperthyroidism	2	300	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.13 [0.04, 0.44]
2 Permanent hypocalcaemia/hypoparathyroidism	2	300	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.04 [0.53, 7.85]
3 Permanent recurrent laryngeal nerve palsy	2	300	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.0 [0.14, 7.15]
4 Regression of Graves' ophthalmopathy	2	260	Odds Ratio (M‐H, Fixed, 95% CI)	1.19 [0.64, 2.21]

2.1. Analysis.

Comparison 2 Total thyroidectomy versus bilateral subtotal thyroidectomy, Outcome 1 Rate of recurrent hyperthyroidism.

Comparison 3. Bilateral subtotal thyroidectomy versus the Dunhill procedure.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Rate of recurrent hyperthyroidism	3	592	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.73 [1.28, 5.85]
2 Permanent hypocalcaemia/hypoparathyroidism	3	592	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.90 [0.47, 7.71]
3 Permanent recurrent laryngeal nerve palsy	3		Peto Odds Ratio (Peto, Fixed, 95% CI)	Totals not selected
4 Regression of Graves' ophthalmopathy	2	90	Odds Ratio (M‐H, Fixed, 95% CI)	0.78 [0.32, 1.92]

3.1. Analysis.

Comparison 3 Bilateral subtotal thyroidectomy versus the Dunhill procedure, Outcome 1 Rate of recurrent hyperthyroidism.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Barczynski 2012.

Methods	Type of study: randomised controlled clinical trial Allocation: 1:1 Intervention model: parallel group superiority design Masking: single‐blinded for participants Primary purpose: treatment of Graves' disease
Participants	Condition: Graves' disease Enrolment: 200 in total Inclusion criteria: patients diagnosed with Graves’ disease and mild active ophthalmopathy undergoing surgical treatment Exclusion criteria: previous thyroid or parathyroid surgery, recurrent hyperthyroidism after radioiodine ablation, a history of Graves’ disease of more than 24 months, thyroid nodules within the posterior aspect(s) of the thyroid lobe(s), suspicion of thyroid cancer, preoperative recurrent laryngeal nerve palsy, pregnancy or lactation, age below 18 years, American Society of Anesthesiologists fitness grade IV, and inability to comply with the follow‐up protocol Diagnostic criteria: clinical features of Graves' disease and TFT supportive of the diagnosis
Interventions	Number of study centres: 1 Treatment before study: patients were rendered euthyroid with thiamazole Titration period: none Intervention(s): bilateral subtotal thyroidectomy Comparator(s): total thyroidectomy
Outcomes	Primary outcome(s): resolution of Graves ophthalmopathy; recurrence of hyperthyroidism Secondary outcome(s): temporary/permanent hypoparathyroidism; temporary/permanent RLN paresis Other outcome(s): ‐
Study details	Run‐in period: none Study terminated before regular end (for benefit/because of adverse events): no Study start date: 2000 Study completion date: 2004
Publication details	Language of publication: English Funding: non‐commercial funding Publication status: peer‐reviewed journal, full article
Stated aim for study	"To evaluate long term results of bilateral subtotal thyroidectomy compared with total thyroidectomy in patients with Graves' disease and mild active ophthalmopathy"
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote from publication: "The randomisation sequence was generated by the computer. Sequencing was based on permutated blocks of two and three to balance the number of patients in the treatment groups" Comment: computer generated random numbers. Does not state if randomisation after consent obtained for surgery; sequence may be guessed easily with use of permutated blocks of two and three
Allocation concealment (selection bias)	Low risk	Quote from publication: "Information on the type of intervention remained in consecutively numbered and sealed envelopes that were stored in the operating theatre. An envelope containing the allocation was added to the patient's file in the operating room. The envelope was opened and the surgeon performed the assigned intervention" Comment: sealed envelope opened by surgeon in theatre
Blinding of participants and personnel (performance bias) Objective outcomes	High risk	Quote from publication: "the patients and ophthalmologist involved in this study were blinded to the group assignment" Comment: patients blinded, surgical team not blinded
Blinding of participants and personnel (performance bias) Subjective outcomes	Low risk	Quote from publication: "the patients and ophthalmologist involved in this study were blinded to the group assignment" Comment: ophthalmology assessment blinded
Blinding of outcome assessment (detection bias) Objective outcomes	Low risk	Quote from publication: "biochemical evaluation consisted of measurement of serum concentrations of TSH, fT3 and fT4 at each visit and serum TBII levels yearly" Comment: objective outcomes assessed by independent third party; thyroid status and calcium status monitored by regular blood tests (externally referenced laboratory)
Blinding of outcome assessment (detection bias) Subjective outcomes	Low risk	Quote from publication: "ophthalmologists involved in this study were blinded to the group assignment" and "indirect laryngoscopy was performed before surgery and on the first day after operation by external ENT specialist" Comment: subjective outcomes assessed by independent third party; (1) ophthalmology assessment of thyroid eye status at regular intervals, (2) RLN palsy checked by external ENT specialist
Incomplete outcome data (attrition bias) Objective outcomes	Low risk	Quote from publication: "CONSORT diagram: five lost to follow up in bilateral subtotal thyroidectomy group, four lost to follow up in total thyroidectomy group" Comment: nine out of 200 lost to follow‐up
Incomplete outcome data (attrition bias) Subjective outcomes	Low risk	Quote from publication: "CONSORT diagram: 5 lost to follow up in bilateral subtotal thyroidectomy group, 4 lost to follow up in total thyroidectomy group" Comment: nine out of 200 lost to follow‐up
Selective reporting (reporting bias)	Low risk	Comment: all outcomes in the methods section reported in the results section (except for mean difference in TSH concentration postoperatively)
Other bias	Low risk	Comment: no other sources of bias detected

Chi 2005.

Methods	Type of study: randomised controlled clinical trial Allocation: 1:1 Intervention model: parallel group superiority design Masking: no blinding of patients or personnel Primary purpose: treatment of Graves' disease
Participants	Condition: Graves' disease Enrolment: 340 in total Inclusion criteria: diagnosis of Graves disease and informed consent to study Exclusion criteria: 1) Previous recurrent hyperthyroidism 2) Large thyroid tumour that prohibited subtotal thyroidectomy 3) Less than 2.5 x 1 x 1 cm thyroid remnant (not suitable for further surgery) 4) Minimum follow‐up period less than 3 months (early loss to follow‐up) Diagnostic criteria: Graves' hyperthyroidism diagnosed with clinical manifestations, TFT, thyroid scan, histology (postoperatively)
Interventions	Number of study centres: 1 Treatment before study: patients were rendered euthyroid with propylthiouracil or methimazole (treatment duration from one month to 144 months) Titration period: none Intervention(s): bilateral subtotal thyroidectomy Comparator(s): unilateral total lobectomy and contralateral subtotal thyroidectomy
Outcomes	Primary outcome(s): 1. Postop thyroid status (TFTs, TSH) Secondary outcome(s): 2. Hypoparathyroidism 3. Recurrent laryngeal nerve injury 4. Wound haematoma/bleeding Other outcome(s): ‐
Study details	Run‐in period: none Study terminated before regular end (for benefit/because of adverse events): no Study start date: 1998 Study completion date: 2002
Publication details	Language of publication: English Funding: not stated Publication status: peer‐reviewed journal, full article
Stated aim for study	To compare the results of bilateral subtotal thyroidectomy and unilateral total and contralateral subtotal thyroidectomy in control of hyperthyroidism in Graves' disease
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote from publication: "patients who underwent subtotal thyroidectomy for hyperthyroidism were prospectively randomised into two groups" Comment: randomisation after consent for surgery; method of sequence generation not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not clearly stated whether allocation concealment took place
Blinding of participants and personnel (performance bias) Objective outcomes	High risk	Quote from publication: "surgery was performed by one surgeon. The standard surgical technique of subtotal thyroidectomy under general anaesthesia with intact inferior thyroid artery was used" Comment: all operations performed by single surgeon; surgeon could not be blinded to group allocation; no blinding of participants
Blinding of participants and personnel (performance bias) Subjective outcomes	High risk	Quote from publication: "surgery was performed by one surgeon. The standard surgical technique of subtotal thyroidectomy under general anaesthesia with intact inferior thyroid artery was used" Comment: all operations performed by single surgeon; surgeon could not be blinded to group allocation; no blinding of participants
Blinding of outcome assessment (detection bias) Objective outcomes	Unclear risk	Quote from publication: " T3 and T4 were measured with radioimmunoassay kit ... total calcium level was measured routinely one day after operation... permanent hypoparathyroidism was defined as symptomatic hypocalcaemia for more than 6 months" Comment: (1) evaluation of thyroid function: laboratory standard tests; (2) operation time: not blinded; (3) blood loss: not blinded; (4) postoperative calcium level: laboratory standard tests; (5) recurrent laryngeal nerve palsy: assessed by laryngoscopy, not clear if blinded
Blinding of outcome assessment (detection bias) Subjective outcomes	Unclear risk	No subjective outcomes are measured in this study
Incomplete outcome data (attrition bias) Objective outcomes	Low risk	Quote from publication: "we excluded patients for whom the follow up period was less than 3 months" Comment: outcome data reported for all 340 participants; however participants lost at an early stage of follow‐up were excluded from study, therefore not intention‐to‐treat analysis
Incomplete outcome data (attrition bias) Subjective outcomes	Low risk	Quote from publication: "we excluded patients for whom the follow up period was less than 3 months" Comment: outcome data reported for all 340 participants; however participants lost at an early stage of follow‐up were excluded from study, therefore not intention‐to‐treat analysis
Selective reporting (reporting bias)	Unclear risk	Comment: participants lost at an early stage of follow‐up were excluded from study, therefore not intention‐to‐treat analysis; mortality not reported
Other bias	Low risk	Comment: none detected

Jarhult 2005.

Methods	Type of study: randomised controlled clinical trial Allocation: 1:1 Intervention model: parallel group superiority design Masking: no blinding of patients or personnel Primary purpose: treatment of Graves' disease
Participants	Condition: Graves' disease Enrolment: 44 in total Inclusion criteria: 1. Diagnosis of Graves disease 2. Preop treatment with antithyroid drugs for at least 3 months 3. Euthyroidism at time of operation 4. Active Graves opthalmopathy of at least moderate degree Exclusion criteria: 1. Previous thyroid or parathyroid surgery 2. Previous treatment with radioiodine 3. Elevation of TSH value more than 200% above normal during antithyroid treatment and within three months prior to surgery Diagnostic criteria: Graves' disease with ophthalmopathy, TFT, TSH antibodies and histological diagnosis based on tissue obtained form surgery.
Interventions	Number of study centres: 10 Treatment before study: preoperative treatment with antithyroid medications for at least 3 months. Titration period: none Intervention(s): total thyroidectomy (complete macroscopic removal of thyroid tissue) Comparator(s): subtotal thyroidectomy (either subtotal bilateral resection or hemithyroidectomy plus unilateral subtotal resection leaving no more than a total of 4 g of remnant tissue)
Outcomes	Primary outcome(s): 1. Change in Graves' ophthalmopathy Secondary outcome(s): 2. Permanent hypoparathyroidism 3. Recurrent laryngeal nerve palsy Other outcome(s): TRAb measurements
Study details	Run‐in period: none Study terminated before regular end (for benefit/because of adverse events): no Study start date: not stated Study completion date: not stated
Publication details	Language of publication: English Funding: non‐commercial funding Publication status: peer‐reviewed journal, full article
Stated aim for study	This study enrolled patients with moderate‐severe endocrine ophthalmopathy and discussed long term results of a prospective, randomised trial of treatment with total or subtotal thyroidectomy
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote from publication: "Forty‐four patients with Graves' disease (GD) and a moderate‐severe Grave's ophthalmopathy were randomly allocated to undergo total or subtotal thyroidectomy" Comment: method of randomisation not specified; took place after consent for surgery
Allocation concealment (selection bias)	Unclear risk	Comment: not stated whether allocation concealment took place
Blinding of participants and personnel (performance bias) Objective outcomes	High risk	Quote from publication: "The design of the study left decisions about the surgical and postoperative care to the attending surgeon" Comment: surgeon could not be blinded to group allocation; note that multiple centres were involved in recruitment for the study
Blinding of participants and personnel (performance bias) Subjective outcomes	High risk	Comment: not stated that patients were blinded to the intervention
Blinding of outcome assessment (detection bias) Objective outcomes	Low risk	Quote from publication: "the follow up consisted of ... biochemical analyses of TSH, free T3 and free T4, TRAB titre and calcium" Comment: evaluation of TRAb, thyroid function and calcium: laboratory standard tests
Blinding of outcome assessment (detection bias) Subjective outcomes	Low risk	Quote from publication: "Evaluation of eye disease activity was carried out by an ophthalmologist... the ophthalmologist was not informed of the type of thyroid surgery performed" Comment: ophthalmopathy ‐ assessed by blinded ophthalmologist
Incomplete outcome data (attrition bias) Objective outcomes	Low risk	Quote from publication: "one patient in the subtotal resection group... refused to come for evaluation after the first two post operative check ups" Comment: one patient lost to follow‐up out of 44; all other outcomes reported
Incomplete outcome data (attrition bias) Subjective outcomes	Low risk	Quote from publication: "one patient in the subtotal resection group... refused to come for evaluation after the first two post operative check ups" Comment: one patient lost to follow‐up out of 44; all other outcomes reported
Selective reporting (reporting bias)	High risk	Comment: postoperative thyroid function (recurrent hyperthyroidism) was measured but not reported; mortality not reported
Other bias	Low risk	Comment: none identified

Muller 2001.

Methods	Type of study: randomised controlled clinical trial Allocation: 1:1 Intervention model: parallel group superiority design Masking: no blinding of patients or personnel Primary purpose: treatment of Graves' disease
Participants	Condition: Graves' disease Enrolment: 152 in total Inclusion criteria: 1. Diagnosis of Graves disease 2. Completed six years follow‐up 3. Euthyroidism at time of operation Exclusion criteria: not specified Diagnostic criteria: hyperthyroidism with endocrine ophthalmopathy or TSH‐R antibodies and characteristics of Graves' disease on ultrasound scan
Interventions	Number of study centres: 1 Treatment before study: preoperative treatment with antithyroid medications if hyperthyroid Titration period: none Intervention(s): 1. Bilateral subtotal thyroidectomy with 3 mL of remaining tissue Comparator(s): 2. Ipsilateral lobectomy and contralateral subtotal resection with 3 mL thyroid remnant per side (6 mL total)
Outcomes	Primary outcome(s): 1. Recurrent hyperthyroidism Secondary outcome(s): 2. Permanent hypoparathyroidism 3. Recurrent laryngeal nerve palsy 4. Wound complications (seroma, infection, haematoma) 5. Progression of Graves' Ophthalmopathy Other outcome(s): ‐
Study details	Run‐in period: none Study terminated before regular end (for benefit/because of adverse events): no Study start date: 1988 Study completion date: 1992
Publication details	Language of publication: English Funding: not stated Publication status: peer‐reviewed journal, full article
Stated aim for study	In this prospective randomised study, the two methods (bilateral subtotal resection and the Dunhill operation) are evaluated with respect to their operative and postoperative complications and their long term results [in treatment of Grave's disease]
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote from publication: "We examined 152 patients... in a prospective randomised study" Comment: the study is prospectively randomised although randomisation techniques not described
Allocation concealment (selection bias)	Unclear risk	Comment: not clearly stated whether allocation concealment took place
Blinding of participants and personnel (performance bias) Objective outcomes	High risk	Quote from publication: "The surgeons were specialised surgeons or were directly supervised by specialised surgeons" Comment: not stated whether surgeons were blinded to group allocation
Blinding of participants and personnel (performance bias) Subjective outcomes	High risk	Comment: not stated whether patients were blinded to group allocation
Blinding of outcome assessment (detection bias) Objective outcomes	Unclear risk	Quote from publication: "Following surgery, patients were started on a thyroxine replacement regimen with iodine and levothyroxine, and the regimen was later adjusted according to their free T3, T4 and TSH levels. All patients had a repeat ... measurement of their plasma calcium" Comment: evaluation of TRAb, thyroid function and calcium: laboratory standard tests
Blinding of outcome assessment (detection bias) Subjective outcomes	Unclear risk	Quote form publication: "the 6 year follow up period ended with a questionnaire sent to the physician currently treating the patient" Comment: unclear whether physician completing follow‐up questionnaire was blinded to intervention; not clear who assessed eye involvement
Incomplete outcome data (attrition bias) Objective outcomes	High risk	Quote from publication: "after 6 years, 99 patients returned the questionnaire" Comment: outcomes only reported in 99/152 (63%) participants
Incomplete outcome data (attrition bias) Subjective outcomes	High risk	Quote from publication: "after 6 years, 99 patients returned the questionnaire" Comment: outcomes only reported in 99/152 (63%) participants
Selective reporting (reporting bias)	Low risk	Comment: outcomes reported appear to be standard for studies addressing this type of question
Other bias	Low risk	Comment: no other risk of bias detected

Witte 2000.

Methods	Type of study: randomised controlled clinical trial Allocation: 1:1:1 Intervention model: parallel group superiority design (three arms) Masking: no blinding of patients or personnel Primary purpose: treatment of Graves' disease
Participants	Condition: Graves' disease Enrolment: 150 in total Inclusion criteria: diagnosis of Graves disease + informed consent to study Exclusion criteria: No diagnosis of Graves disease Does not consent to study Not medically fit for surgical management Diagnostic criteria: clinical features of Graves' disease with TSH‐R antibodies and hyperthyroidism
Interventions	Number of study centres: 1 Treatment before study: preoperative treatment with antithyroid medications, beta‐blockers and thyroid hormone substitution. Titration period: none Intervention(s): total thyroidectomy Comparator(s): Dunhill procedure (total lobectomy + contralateral subtotal thyroidectomy) Bilateral subtotal thyroidectomy
Outcomes	Primary outcome(s): Graves' ophthalmopathy Postoperative TSH‐R antibody level Secondary outcome(s): RLN palsy Hypoparathyroidism Other outcome(s): Wound complications/infection rate
Study details	Run‐in period: none Study terminated before regular end (for benefit/because of adverse events): no Study start date: 1993 Study completion date: 1995
Publication details	Language of publication: English Funding: not stated Publication status: peer‐reviewed journal, full article
Stated aim for study	We investigated whether total thyroidectomy might be superior to subtotal thyroidectomy (bilateral or unilateral remnants of less than 4 mL) in respect to the adverse effect on immunologic parameters and endocrine orbitopathy
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote from publication: "After the patients consented to the prospective randomised trial, 50 were randomised to total thyroidectomy, 50 to unilateral total and contralateral subtotal thyroid resection, and 50 to bilateral subtotal thyroid resection" Comment: randomisation after consent for surgery; method of randomisation not specified
Allocation concealment (selection bias)	Unclear risk	Comment: not reported whether allocation concealment took place
Blinding of participants and personnel (performance bias) Objective outcomes	High risk	Comment: not stated that surgeons were blinded to group allocation
Blinding of participants and personnel (performance bias) Subjective outcomes	High risk	Quote from publication: "in case of specific technical demands, surgeons were allowed to change to recommended procedure individually for the patient's advantage" Comment: in some cases group allocation changed intra‐operatively for technical reasons
Blinding of outcome assessment (detection bias) Objective outcomes	Unclear risk	Quote from publication: "pre‐operative and post operative TSH‐Ab titres were measured"; "All patients were checked by certified ear, nose and throat physicians for recurrent laryngeal nerve paralysis" Comment: thyroid status and calcium status were assessed with standardised referenced assays; recurrent laryngeal nerve status was assessed by ENT specialty review, unclear whether blinded
Blinding of outcome assessment (detection bias) Subjective outcomes	Unclear risk	Quote from publication: "GO was estimated clinically, using the modified scale system of the American Thyroid Association, by general practitioners, ophthalmologists and by the patients themselves" Comment: Graves' ophthalmopathy was assessed by ophthalmologist and patient; unclear whether blinding took place
Incomplete outcome data (attrition bias) Objective outcomes	Unclear risk	Quote from publication: "the TSH‐R antibodies were preoperatively measured in 128 of the 150 patients... post operative TSH‐R‐Abs were measured in 107 of the 150 patients" Comment: preoperative/postoperative TSH‐R antibodies not measured in 14.7% and 28.7% of participants; GO/TSH‐R antibodies/long term calcium all measured from > 18‐58 months postoperatively
Incomplete outcome data (attrition bias) Subjective outcomes	Unclear risk	Quote from publication: "GO was estimated clinically, using the modified scale system of the American thyroid association, by general practitioners, ophthalmologists and by the patients themselves" Comment: all Graves' ophthalmopathy patients underwent a standardised assessment pre‐ and postoperatively using ATA criteria
Selective reporting (reporting bias)	Unclear risk	Quote from publication: "the TSH‐R antibodies were preoperatively measured in 128 of the 150 patients... post operative TSH‐R‐Abs were measured in 107 of the 150 patients" Comment: a large number of participants had no assessment of the TSH‐R antibodies postoperatively
Other bias	Low risk	Comment: no other risk of bias detected

ATA: American Thyroid Association; ENT: ear, nose and throat; fT3: free tri‐iodothyronine; fT4: free thyroxine; GO: Graves' opthalmopathy; RLN: recurrent laryngeal nerve; TBII: thyrotropin binding inhibiting immunoglobulins; TFT: thyroid function tests; TRAb: anti‐thyroid receptor antibody; TSH: thyroid‐stimulating hormone; TSH‐R: thyroid‐stimulating hormone receptor

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abe 1998	Non randomised, no surgical comparison arm
Al‐Adhami 2012	Retrospective study
Barakate 2002	Retrospective study
Boostrom 2007	Prospective study, non‐randomised
Campos 1994	Only one surgical arm
Candela 2006	Retrospective case series
De Bellis 2012	Non randomised, no surgical comparison arm
Fernandez‐Sanchez 1993	Only one surgical arm
Feroci 2014	Review
Guo 2013	Review
Hughes 2011	Total thyroidectomy both arms, non‐randomised
Hussain 2008	Non‐randomised, no surgical comparison arm
Jarhult 2012	Non randomised
Jukic 2010	Radio‐iodine versus surgery, no surgical comparison arm
Khandra 1992	Retrospective study
Ku 2005	Historical control group, non‐randomised
Kurihara 2002	Retrospective study
Leo 2012	No surgical comparison arm
Lepner 2008	Prospective study, non‐randomised
Ljunggren 1998	No surgical comparison arm
Lofthouse 2006	Review
Marcocci 1992	Retrospective study, no surgical comparison arm
Maschuw 2011	Protocol only, recruitment not complete due to lack of funding
Menconi 2007	Subtotal thyroidectomy in both arms, no comparison of alternative surgical technique
Miccoli 1996	Non‐randomised
Morosini 1988	Retrospective study
Robert 2001	Retrospective study
Rosato 2002	Retrospective study
Sanchez 1999	Retrospective study
Sugino 2008	Non‐randomised, no comparison of alternative surgical techniques
Tallstedt 1992	No alternative surgical technique comparison
Tallstedt 1997	Only one surgical arm
Tian 2012	No surgical comparison arm
Torring 1996	Only one surgical arm
Welch 2011	Consecutive, non‐randomised participants
Woods 2014	Meta‐analysis
Yalcin 2010	Retrospective study
Zaraca 2000	Retrospective study

Characteristics of ongoing studies [ordered by study ID]

Maschuw 2012.

Trial name or title	Total versus near‐total thyroidectomy in Graves' disease and their outcome on postoperative transient hypoparathyroidism: study protocol for a randomised controlled trial
Methods	Type of study: randomised controlled clinical trial Allocation: 1:1 Intervention model: parallel group superiority Masking: blinding of patients and observer Primary purpose: to compare rates of transient postoperative hypoparathyroidism in total and subtotal thyroidectomy
Participants	Condition: Graves' disease, for definite surgery Inclusion criteria: diagnosis of Graves disease and informed consent to study Exclusion criteria: conservative treatment, malignancy, previous thyroid surgery, coincident hypoparathyroidism
Interventions	Intervention: bilateral subtotal thyroidectomy, remnants of < 1mL on each side Control: total thyroidectomy
Outcomes	Primary outcome: transient postoperative hypoparathyroidism Secondary outcomes: reoperations due to bleeding, recurrent laryngeal nerve palsy, permanent hypoparathyroidism, recurrent disease, changes of endocrine orbitopathy, quality of life within a one‐year of follow‐up period
Starting date	Recruitment not started
Contact information	maschuw@med.uni‐marburg.de
Study identifier	German clinical trials register (DRKS) DRKS00004161
Official title	Total versus near‐total thyroidectomy in Graves' disease and their outcome on postoperative transient hypoparathyroidism: study protocol for a randomised controlled trial
Stated purpose of study	The trial is designed as a prospective randomised controlled patient and observer blinded multi‐centred trial in a parallel design including an active comparator and an intervention group. The intervention addresses the surgical procedure: near‐total thyroidectomy leaving bilateral remnants of <= 1 mL on each side in the intervention group and total thyroidectomy in the control group. The occurrence of transient postoperative hypoparathyroidism is defined as primary endpoint
Notes

Differences between protocol and review

Originally we planned to only investigate total versus subtotal thyroidectomy for Graves' disease and Graves' ophthalmopathy. However, to provide maximum information for patients, clinicians and decision makers we also included all types of thyroidectomy for these conditions, such as bilateral thyroidectomy versus unilateral total and contralateral subtotal thyroidectomy (Dunhill procedure). Therefore, the title was changed from ‘Total thyroidectomy versus subtotal thyroidectomy for Graves' disease’ to 'Thyroid surgery for Graves’ disease and Graves’ ophthalmopathy'.

We were not able to conduct analyses of health‐related quality of life or socioeconomic effects (health economic outcomes were changed to the broader term socioeconomic effects) related to thyroidectomy in Graves' disease due to insufficient data from the included studies. We were also not able to conduct a meta‐analysis of some postoperative complications (e.g. bleeding/wound infection), as reporting criteria were not sufficiently similar between studies.

We inserted a new paragraph 'Quality of evidence' under 'Data synthesis' in the 'Methods' section.

Contributions of authors

All review authors read and approved the final review draft.

Zi Wei Liu (ZWL): protocol draft, search strategy development, acquiring trial copies, trial selection, data extraction, data analysis, data interpretation, and future review update.

Liam Masterson (LM): protocol draft, search strategy development, acquiring of trial copies, trial selection, data extraction, data analysis, data interpretation, and future review update.

Brian Fish (BF): protocol draft, trial selection, data interpretation, and future review update.

Piyush Jani (PJ): protocol draft, trial selection, data interpretation, and future review update.

Krishna Chatterjee (KC): protocol draft, trial selection, data interpretation, and future review update.

Declarations of interest

Zi Wei Liu: none known.

Liam Masterson: none known.

Piyush Jani: none known.

Brian Fish: none known.

Krishna Chatterjee: none known.

New