Total thyroidectomy versus subtotal thyroidectomy in treatment of multinodular goiter: a meta-analysis

Background The surgical management of multinodular goiter is the treatment of choice for the majority of cases. There is controversy between radical resection with the lifelong thyroxine substitution and function-preserving resection with the risk for recurrence, and the complications associated with total thyroidectomy. Objective The aim of the present study was to compare the outcome of total thyroidectomy (TT) in comparison with subtotal thyroidectomy (ST) as regards hypocalcemia, transient nerve injury, and recurrence rate. Patients and methods This study included published English medical articles in the last 20 years, concerning the treatment of multinodular goiter. Results Meta-analysis was for the evaluation of surgical outcomes after surgical management of multinodular goiter including total thyroidectomy versus subtotal thyroidectomy by comparing the TT versus ST in our study including 23 studies included of the total number of patients (4485) who underwent subtotal thyroidectomy versus the total number of patients (7116) who underwent total thyroidectomy; a comparison was done as regards postoperative complications including RLN injury (transient or permanent), rate of recurrence, need for reoperation, and incidence of postoperative hypocalcemia. The incidence of RLN injury is lower in ST compared to TT, and its recurrence is much lower in TT than in ST. Conclusion This meta-analysis showed that the advantages of total thyroidectomy include adequate eradication of the disease, prevention of recurrent goiter, and avoidance of the need for completion surgery in case of occult malignancy, but it is associated with higher morbidity (postoperative thyroidectomy complications: RLN palsy and hypoparathyroidism) and the need for lifelong replacement therapy (L-thyroxin supplementation).


Patients and methods
The study was done in the following steps: • Determination of the target disease.
• Identification and location of articles.
• Screening and evaluation of articles.• Data collection.• Data analysis.
• Discussion and conclusion.

Target subject
The target is to compare two techniques of thyroidectomy in the treatment of multinodular goiter, total thyroidectomy, and subtotal thyroidectomy as regards postoperative complications and recurrence.

Identification and location of articles
Studies included published medical articles about the advantages and disadvantages of each total thyroidectomy and subtotal thyroidectomy through searching the MEDLINE database (www.pubmed.com) using the following keywords in different combinations:

• Total thyroidectomy • Subtotal thyroidectomy • Multinodular goiter • Postoperative complication • Multinodular goiter recurrence • Incidental thyroid cancer
Over 8000 articles were found.After customizing the date, age, and language, they narrowed to about 3796 articles, and after the exclusion of non-relevant articles of 3780, about 16 relevant articles were found.By application of inclusion criteria, 6 articles were found meeting the inclusion criteria and can undergo meta-analysis.

Screening and evaluation
The screening form of articles was used by the investigators to screen the articles, which were yielded by the MEDLINE search after blinding the author name and journal name.The screen form of the articles is as follows: 1) Irrelevant articles: articles that may have one of the keywords but a different purpose from our study (3780).
2) Relevant articles: after the exclusion of repeated and non-relevant articles, articles that contain one or more of the above keywords (16).3) Included articles: These are 6 articles that fulfilled the following inclusion criteria and were included for further steps of data collection, analysis, and reporting (Table 1): • Published in the English language.
• Conducted on human subjects.
• Adults more than 18 years • All patients are euthyroid.
• No patients with suspicion of thyroid malignancy.
• Only randomized clinical trial, retrospective, and prospective studies.• No articles with different outcome measures.4) Excluded articles: Articles that miss one or more of the above mentioned inclusion criteria (10).The following table will show the excluded articles and the reason for exclusion Table 2:

Data collection
Information was gathered for each individual study on total thyroidectomy and subtotal thyroidectomy in the treatment of multinodular goiter, and we extracted data from them (postoperative complications and recurrence).

Data analysis Statistical methods
Statistical analysis was done using Comprehensive Meta Analysis© version 2 (Biostat ™ , NJ, USA).

Testing for heterogeneity
Studies included in the meta-analysis were tested for heterogeneity of the estimates using the following tests: 1. Cochran Q chi-square test: A statistically significant test (p value < 0.1) denoted heterogeneity among the studies.2. I-squared (I2) index which is calculated as follows: The I-squared is interpreted as follows: ■ 0 to 40%: might not be important ■ 30 to 60%: may represent moderate heterogeneity ■ 50 to 90%: may represent substantial heterogeneity ■ 75 to 100%: considerable heterogeneity

Effect size estimation
The effect size for binary outcomes was expressed as relative risk (RR) with its 95% confidence limits (95% CI).

Pooling of estimates
Estimates from included studies were pooled using both the Mantel-Haenszel fixed-effects method (FEM) and the DerSimonia Laird random-effects method (REM).In the absence of significant heterogeneity, the FEM was considered; otherwise, the REM was considered.

Examination of publication bias
Publication bias was assessed by examination of funnel plots.A funnel plot is a plot of the estimated effect size (RR) on the horizontal axis versus the standard error (SE) for the effect size as a measure of study size on the vertical axis.
Large studies appear toward the top of the graph and tend to cluster near the mean effect size.Smaller studies appear toward the bottom of the graph and (since there is more sampling variation in effect size estimates in the smaller studies) will be dispersed across a range of values.
In the absence of publication bias, the studies are expected to be distributed symmetrically about the combined effect size.By contrast, in the presence of bias, it is expected that the bottom of the plot would show a higher concentration of studies on one side of the mean than the other.This would reflect the fact that smaller studies (which appear toward the bottom) are more likely to be published if they have larger-than-average effects, which makes them more likely to meet the criterion for statistical significance.

Level of significance
A two-sided p value < 0.05 denoted statistical significance.

Meta-analysis of the overall incidence of RLN palsy
There is no evidence for heterogeneity among the included studies (Q = 4.887, DF = 4, P value 0.299, I 2 = 18.14%) pooling of the studies using fixed effects showed a relative risk of 1.6 with 95% CI of 1.27 to 2.02 which was statistically significant (> 0.001).
Table 3 shows the results of the meta-analysis of the overall incidence of RLN palsy.
Pooling of the studies using fixed effects showed a relative risk of 1.6 with a 95% CI of 1.27 to 2.02 which was statistically significant (> 0.001).

Meta-analysis of the incidence of permanent RLN palsy
No evidence for heterogeneity among the included studies (Q = 4.887, DF = 4, P value 0.2991, I 2 = 18.14%) pooling of the studies using fixed effects showed a relative risk of 1.31 with 95% CI of 0.91 to 1.89 which was statistically not significant (p value 0.149).
Table 4 shows the results of the meta-analysis of the incidence of permanent RLN palsy.
Pooling of the studies using fixed effects showed a relative risk of 1.31 with a 95% CI of 0.91 to 1.89 which was statistically not significant (p value 0.149).
A funnel plot for the incidence of permanent RLN palsy was done.There is evidence for publication bias.

Meta-analysis of incidence of temporary RLN palsy
No evidence for heterogeneity among the included studies (Q = 7.060, DF = 4, P value 0.133, I 2 = 43.34%)pooling of the studies using fixed effects showed a relative risk of 1.828 with 95% CI of 1.36 to 2.46 which was statistically significant (p value > 0.001).
Table 5 shows the results of the meta-analysis of the incidence of temporary RLN palsy.
Pooling of the studies using fixed effects showed a relative risk of 1.828 with a 95% CI of 1.36 to 2.46 which was statistically significant (p value > 0.001).
A funnel plot for the incidence of temporary RLN palsy was done.There is evidence of publication bias.

Meta-analysis of the overall incidence of hypoparathyroidism
There is evidence for heterogeneity the included studies (Q = 66.065,DF = 4, P value > 0.0001, I 2 = 93.95%)pooling of the studies using random effects showed a relative risk of 1.30 with 95% CI of 1.20 to 1.41 which was statistically significant (p value 0.036).
Table 6 shows the results of the meta-analysis of the overall incidence of hypoparathyroidism.
There is evidence for heterogeneity among the included studies (Q = 66.065,DF = 4, P value > 0.0001, I 2 = 93.95%).Pooling of the studies using random effects showed a relative risk of 1.30 with a 95% CI of 1.20 to 1.41 which was statistically significant (p value 0.036).
A funnel plot for the overall incidence of hypoparathyroidism was done.There is evidence of publication bias.

Meta-analysis of the incidence of permanent hypoparathyroidism
No evidence for heterogeneity among the included studies (Q = 0.883, DF = 4, P value 0.927, I 2 = 0.00%) pooling of the studies using fixed effects showed a relative risk of 1.48 with 95% CI of 1.11 to 1.96 which was statistically significant (p value 0.007).
Table 7 shows the results of a meta-analysis of the incidence of permanent hypoparathyroidism.
Pooling of the studies using fixed effects showed a relative risk of 1.48 with a 95% CI of 1.11 to 1.96 which was statistically significant (p value 0.007).
A funnel plot for the incidence of permanent hypoparathyroidism was done.There is evidence of publication bias.

Meta-analysis of Incidence of temporary hypoparathyroidism
There is evidence for heterogeneity among the included studies (Q = 66.517,DF = 4, P value > 0.0001, I 2 = 93.99%)pooling of the studies using random effects showed a relative risk of 2.389 with 95% CI of 1.02 to 5.61 which was statistically significant (p value 0.046).Table 8 shows the results of the meta-analysis of the incidence of temporary hypoparathyroidism.
Pooling of the studies using random effects showed a relative risk of 2.389 with a 95% CI of 1.02 to 5.61 which was statistically significant (p value 0.046).
A funnel plot for the incidence of temporary hypoparathyroidism was done.There is evidence of publication bias.

Meta-analysis of the incidence of recurrence
There is evidence for heterogeneity among the included studies (Q = 13.484,DF = 4, P value 0.009, I 2 = 70.34%)pooling of the studies using random effects showed a relative risk of 0.013 with 95% CI of 0.002 to 0.11 which was a statistically significant p value of > 0.001.
Table 9 shows the results of the meta-analysis of the incidence of recurrence.
Pooling of the studies using random effects showed a relative risk of 0.013 with a 95% CI of 0.002 to 0.11 which was statistically significant (p value > 0.001).
A funnel plot for the incidence of recurrence was done.There is evidence of publication bias.

Discussion
Surgical management is considered the treatment of choice for multinodular goiter [19].Several surgical procedures for MNG, including subtotal thyroidectomy and   [20,21], and more recent studies have shown comparison of both surgical approaches regarding the postoperative complications [22].Increasing numbers of total thyroidectomies are being performed, and the indications for this procedure include thyroid cancer, Graves' disease, and multinodular goiter [23].As it completely eradicates the disease process and lowers the local recurrence rate, it avoids the need for completion thyroidectomy with minimal risk of morbidity [24].
In this study, we compare the outcome of total thyroidectomy in comparison with subtotal thyroidectomy as regards hypocalcemia, hypothyroidism, recurrence rate, and RLN injury.
As regards RLN palsy, there is a higher incidence in total over subtotal in temporary and permanent RLN palsy with a P value < 0.001 which is statistically significant.
As regards postoperative hypoparathyroidism, there is a higher incidence in total over subtotal in postoperative hypoparathyroidism whether it is temporary or permanent, with a P value < 0.001 which is statistically significant.
As regards recurrence, there is a higher incidence in subtotal over the total in the overall incidence of recurrence with a P value < 0.001 which is statistically significant.
Ozbas et al. reported that the incidence rate of recurrence of MNG after bilateral subtotal thyroidectomy ranges between 1.2% [16], Delbridge reported that subtotal thyroidectomy for benign thyroid disease has been performed for more than a century and that it may reduce the associated risk of postoperative hypocalcemia and recurrent laryngeal nerve (RLN) palsy [25].Most surgeons still argue whether the potential risk of total thyroidectomy (TT) outweighs the potential benefits [26].
Regarding the RLN injury postoperative either transient or permanent our study showed a statistically significant difference between subtotal thyroidectomy over total thyroidectomy as in subtotal thyroidectomy, we avoid dissection over RLN or using excessive diathermy also keep in mind most of them are neuropraxia, nerve traction, or edema resolving by the time.
Among included studies, the results showed that subtotal thyroidectomy has improved transient hypothyroidism and permanent hypothyroidism over total thyroidectomy with preservation of the posterolateral part to avoid postoperative hypothyroidism and the need for replacement therapy.
TT has become a preferred surgical procedure for MNG for the majority of surgeons because it eliminates the risk of recurrence, and there is no need for reoperation [26].A lot of data has been published indicating the equal incidence of both transient/permanent RLN palsy and transient/permanent hypocalcemia for subtotal thyroidectomy and total thyroidectomy [27].
Regarding the recurrence rate, the subtotal thyroidectomy showed less significance than total thyroidectomy as total thyroidectomy ablates any remaining tissue [28].

Conclusion
This meta-analysis showed that the advantages of total thyroidectomy include adequate eradication of the disease, prevention of recurrent goiter, and avoidance of the

Table 1
Included articles

Table 2
Excluded articles and reason for exclusion

Table 3
Meta-analysis of the overall incidence of RLN palsy

Table 4
Meta-analysis of incidence of permanent RLN palsy

Table 5
A meta-analysis of the incidence of temporary RLN palsy

Table 6
Meta-analysis of the overall incidence of hypoparathyroidism

Table 7
Meta-analysis of the incidence of permanent hypoparathyroidism

Table 8
Meta-analysis of the incidence of temporary hypoparathyroidism

Table 9
Meta-analysis of the incidence of recurrence