Systematic review of cure and recurrence rates following minimally invasive parathyroidectomy

Background The majority of patients with primary hyperparathyroidism (PHPT) have a single overactive adenoma. Advances in preoperative imaging and surgical adjuncts have given rise to minimally invasive parathyroidectomy (MIP), with lower complication rates in comparison with bilateral neck exploration. Misdiagnosis and undertreatment of multiglandular disease, leading to potentially higher recurrence rates, remains a concern. This study evaluated risks of long‐term (1 year or more) recurrence following ‘targeted’ MIP in PHPT. Methods Multiple databases were searched for studies published between January 2004 and March 2017, looking at long‐term outcomes (1 year or more) following targeted MIP for PHPT. English‐language studies, with at least 50 patients and a mean follow‐up of 1 year, were included. Results A total of 5282 patients from 14 studies were included. Overall mean recurrence and cure rates were 1·6 (range 0–3·5) and 96·9 (95·5–100) per cent respectively. Mean follow‐up was 33·5 (1–145) months. When intraoperative parathyroid hormone (PTH) measurements were not done, cure rates were higher (99·3 per cent versus 98·1 per cent with use of intraoperative PTH measurement; P < 0·001) and recurrence rates lower (0·2 versus 1·5 per cent respectively; P < 0·001). Conclusion Targeted MIP for a presumed single overactive adenoma was associated with very low recurrence rates, without the need for intraoperative PTH measurement when preoperative imaging studies were concordant. Targeted MIP should be encouraged.


Introduction
Primary hyperparathyroidism (PHPT) is a common condition, with an estimated incidence of one to seven per 1000 adults 1 . The condition is detected incidentally in more than 80 per cent of subjects on routine biochemical analysis 2,3 . Surgery remains the only curative option.
Parathyroid surgery via an open bilateral four-gland neck exploration was first performed in 1925 4 and remained the standard treatment until the early part of the 21st century. In experienced hands, this method has a cure rate of at least 95 per cent, with a morbidity rate of less than 3 per cent 5 , and does not require any form of preoperative localization imaging.
It is widely recognized that in over 85 per cent of patients with PHPT the cause is a single overactive parathyroid adenoma, often identifiable through preoperative imaging, allowing selective removal 5 . The two most common modalities of preoperative localization are sestamibi imaging and high-resolution ultrasonography. A meta-analysis 6 examining the value of ultrasound imaging in PHPT found an overall pooled sensitivity of 76⋅1 per cent and a positive predictive value of 93⋅2 per cent, although operator and centre variation was acknowledged 7 . With negative localization of a solitary adenoma, the likelihood of multiglandular disease is reported to be up to 30 per cent 8,9 .
Surgical management of PHPT has evolved over the past 20 years. Rapid intraoperative parathyroid hormone (IOPTH) assay 10 , sestamibi scintigraphy and radio-guided parathyroidectomy 11  Targeted minimally invasive parathyroidectomy (MIP) has been shown in large studies, systematic reviews and meta-analysis to be highly effective with low complication rates compared with open bilateral neck exploration (BNE) 14 -16 . Advantages of MIP include shorter duration of surgery, lower rates of postoperative hypocalcaemia, less postoperative pain and a smaller scar 17 . There are also potential financial advantages in performing MIP 18,19 , with lower operative costs and more rapid hospital discharge.
The consensus statement published by the European Society of Endocrine Surgeons (ESES) 20 described MIP as a safe and cost-effective procedure for the treatment of selected patients with PHPT. In the UK, the National Institute for Health and Care Excellence (NICE) released guidelines for minimally invasive video-assisted parathyroidectomy (MIVAP) 21 , stating that 'current evidence on the efficacy and safety of MIVAP is adequate to support the use of this procedure'.
Misdiagnosis and the risk of undertreating multigland disease, leading to high recurrence rates, nevertheless remains a potential shortcoming of this approach. This systematic review aimed to examine long-term (1 year or more) recurrence rates following targeted MIP in PHPT.

Acquisition of evidence
The PRISMA protocol 22 was followed to perform a comprehensive literature search using MEDLINE, Embase, CINAHL, the UK Clinical Trials Gateway and the US Trials Database between January 2004 and March 2017. The PICO framework 23 was used and terms combined with Boolean operators (AND, OR) to refine the search further.
Two independent reviewers identified all studies that met the inclusion criteria for full review. References of the searched studies were evaluated for potential inclusion in the review. Where possible, contact was attempted with To analyse the best available data focusing on long-term (follow-up of at least 1 year) recurrence and cure rates, the inclusion criterion for centres performing MIP regularly (at least 50 patients per annum) was set. This was done to minimize skewing of data from 'low-volume' centres and small case series.

Study quality and levels of evidence
The quality of studies and risk of bias were assessed by two reviewers. All studies included in the review were non-randomized, and therefore the Methodological Index for Non-Randomized Studies (MINORS) tool 24 was used. This tool assessed non-randomized studies on the following criteria: clearly stated aims, inclusion of consecutive patients, prospective data collection, appropriate endpoints, unbiased evaluation of endpoints, appropriate duration of follow-up and loss to follow-up no more than 5 per cent. For comparative studies, further criteria were assessed: whether the control group underwent a standard intervention, the use of contemporary groups, baseline equivalence of the groups, prospective calculation of the sample size and statistical analysis adapted to the study design.
To ascertain the level of evidence of the included studies, the Oxford Centre for Evidence-based Medicine (OCEM) 25 guidelines were employed.

Outcomes
The primary outcome was recurrence rate, defined as the rate of hypercalcaemia occurring after 6 months of normocalcaemia following parathyroidectomy.
Secondary outcomes were: cure rate (defined as normocalcaemia persisting for more than 6 months after surgery), type of MIP performed, use of IOPTH measurements and postoperative complication rate.
When analysing IOPTH use and non-use during MIP, the studies were divided into those that included only patients who exhibited positive preoperative concordant imaging (PC cohort) and those that included a heterogeneous cohort (H cohort: 1-2 image-positive or image-negative patients).

Statistical analysis
GraphPad Prism ® version 7.0d (GraphPad Software, La Jolla, California, USA) was used for statistical analysis, and Microsoft Excel ® version 16.12 (Microsoft, Redmond, Washington, USA) for data handling. The statistical significance of categorical variables was determined with Fisher's exact test. P < 0⋅050 was considered statistically significant.

Study identification
The initial literature search yielded a total of 252 studies, of which 14 11,20,26 -37 met the inclusion criteria (Fig. 1). All studies were observational and included a total of 5282 patients who had targeted MIP for PHPT. There was a female preponderance of 3⋅4 : 1 and the overall mean age was 58⋅9 years ( Table 1).
All studies defined cure as normalization of serum calcium levels, and all recorded serum calcium levels at each follow-up appointment. All but three studies 26,31,37 also measured postoperative serum PTH levels routinely.
All studies used the widely accepted definition of 'disease recurrence' as hypercalcaemia after 6 months of proven normocalcaemia following initial surgery.
The most common type of surgery performed was the open MIP technique 26,27,29 -32,34-37 , but other approaches included MIVAP 20,28 , minimally invasive radio-guided parathyroidectomy 11 and a totally endoscopic MIP technique 33 . When reported, studies had incision lengths of less than 3 cm, whereas studies that did not report on incision length stated that their procedure was 'minimally invasive' or a 'focused/lateral exploration'.
Four studies 11,29,31,32 with a total of 2908 patients performed targeted MIP without IOPTH measurements, and nine studies 20,26 -28,30,33-36 carried out targeted MIP with IOPTH measurements in 2072 of the patients. One study 37 made no reference to the use of this test, so was not included in the analysis. Five studies 28 -32 included only patients with positive, concordant preoperative localization studies on at least two different imaging modalities, seven 11,20,27,33 -36 included patients regardless of imaging findings, and two 26,37 did not report on imaging findings.
An overall complication rate of 4⋅4 per cent in 4010 patients was reported in nine studies 20,27 -34 , whereas five 11,26,35 -37 did not report complication rates. Transient and permanent postoperative hypocalcaemia rates were 1⋅6 and 0⋅05 per cent respectively. Temporary recurrent laryngeal nerve (RLN) palsy occurred in 1⋅1 per cent and permanent RLN palsy in 0⋅3 per cent of the 4010 patients. Complications are summarized in Table 3.
There were significant differences in how complications were defined and reported. In terms of RLN palsy, four studies 20,28,31,40 routinely used laryngoscopy after surgery, one study 29 used laryngoscopy in patients with postoperative dysphonia, and another study 30 explicitly stated that laryngoscopy was not used. The remaining publications did not comment on postoperative laryngoscopy. The definition of hypocalcaemia also varied, from a serum calcium level lower than 2 mmol/l 28 to 1⋅95 mmol/l or less 30 , whereas other studies did not define cut-off levels.

Discussion
This systematic review has indicated that, despite variations in technique, targeted MIP was associated with low long-term (at least 1 year) recurrence rates of only about 1⋅5 per cent. Overall recurrence rates were lowest and cure rates highest in studies that included patients where there was positive, concordant preoperative localization. Interestingly, the analysis suggested that recurrence and cure rates were better when IOPTH measurements were not used. As indicated in Table 2, this appeared to reflect use of MIP without routine IOPTH measurement in patients with concordant imaging. IOPTH measurement seemed to be used more frequently in non-concordant or image-negative patients.
Two reports 38,39 relating to MIP were not included in this systematic review as they did not fulfil the inclusion criteria (mean duration of follow-up not documented). These studies reported a median follow-up of 9 (range 0-116) months. The first study 38 investigated whether MIP was associated with a higher recurrence rate than BNE, and the second 39 considered variables that might predict recurrence in parathyroidectomy for PHPT. In both studies, Kaplan-Meier curves were constructed to determine disease-free estimates for MIP and BNE. No statistically significant differences were found between MIP and open parathyroidectomy in either study (P = 0⋅55 and P = 0⋅59 respectively).
This review also demonstrated the overall complication rate to be low (4⋅4 per cent in 4010 patients), and similar to values reported by previous large studies 15,40 involving BNE, together with similar rates of permanent and temporary recurrent laryngeal nerve palsy (0⋅3 and 1⋅1 per cent respectively). The lack of postoperative laryngoscopy in the present review and in other studies, however, may mean that the true rate of recurrent nerve palsy is underestimated. Similar consideration applies to rates of postoperative hypocalcaemia, owing to the different cut-off levels. This systematic review analysed 14 level 2b studies with a median MINORS score of 10 (mean 9⋅8), indicating evidence of fair quality. However, its main limitation was the lack of uniformity in the presentation and reporting of data by the individual studies. This is a widely recognized and inherent problem of collating data from observational studies.
Despite these shortcomings, targeted MIP in the surgical management of patients with a presumed single overactive parathyroid adenoma is a safe technique that provides long-term cure. The review also suggests that MIP is effective without intraoperative hormone estimations in patients with preoperative concordant imaging.

Disclosure
The authors declare no conflict of interest.