Causes of Emergency Department Visits Following Thyroid and Parathyroid Surgery

Importance  With reimbursement being increasingly tied to outcome measures, minimizing unexpected health care needs in the postoperative period is essential. This article describes reasons for emergency department (ED) evaluation, rates of readmission to the hospital, and significant risk factors for readmission during the postoperative period.

Objective  To describe the subset of patients requiring ED evaluation within 30 days of thyroidectomy or parathyroidectomy and their associated risk factors.

Design, Setting, and Patients  Retrospective chart review in a tertiary care center of adult patients who underwent thyroidectomy or parathyroidectomy between January 1, 2009 and October 7, 2010. Patients were identified from an institutional review board–approved database. Postoperative patients who visited the emergency department (ED) within the first 30 days following surgery were selected and compared with the postoperative patients who did not visit the ED.

Exposures  Thyroidectomy or parathyroidectomy.

Main Outcomes and Measures  Statistical analysis evaluated the association of demographic and clinical characteristics between the patients who required ER evaluation and those who did not. Clinical characteristics evaluated included type of surgery, medical comorbidities, and proton pump inhibitor (PPI) usage. Multiple logistic regression predicted the odds of an ED visit based on presence of diabetes, gastroesophageal reflux disease (GERD), or PPI use.

Results  Of the 570 patients identified, 64 patients required a visit to the ER a total of 75 times for issues including paresthesias (n = 28), wound complications (n = 8), and weakness (n = 6). Fifteen hospital admissions occurred for treatment of a variety of postoperative complications. A significant association was found between the presence of diabetes (P = .03), GERD (P = .04), and the current use of PPIs (P = .03). When controlling for diabetes and GERD, we found that patients taking PPIs were 1.81 times more likely to visit the ED than patients not taking PPIs (P = .04).

Conclusions and Relevance  Patients taking PPIs are 1.81 times more likely to require ED evaluation than those who are not taking PPIs.

Perioperative complications from thyroidectomy and parathyroidectomy are well described.^1-4 Recurrent laryngeal nerve injury with its associated voice and airway issues, electrolyte abnormalities, and wound infections are among the many issues that can arise following surgery. These issues contribute to the overall cost of patient care but are not always present in the immediate postoperative period. This is especially relevant given the increasing discussion of outpatient head and neck endocrine surgery.^5,6 Missing among the literature is a description of the postoperative issues requiring an emergency department (ED) evaluation in a large collection of patients who have undergone thyroid or parathyroid surgery.^7,8 The present study describes risk factors that increase the likelihood for ED evaluation within 30 postoperative days following thyroidectomy and parathyroidectomy in a tertiary health care center, and it also reports some of the ED findings in this group of patients.

Patients undergoing thyroidectomy or parathyroidectomy at a tertiary care center between January 1, 2009, and October 7, 2010, were identified via an institutional review board–approved thyroid/parathyroid database. Medical records in this database were reviewed to identify postoperative patients who visited the Henry Ford Health System ED within the first 30 days following surgery as well as those who did not. Demographic data points collected included age, sex, body mass index (BMI), and race. Clinical characteristics evaluated included type of surgery, proton pump inhibitor (PPI) use, and medical comorbidities such as diabetes mellitus (DM), hypertension (HTN), end-stage renal disease (ESRD), and gastroesophageal reflux disease (GERD). Timing of the ED visit with respect to the date of surgery was noted. Diagnoses identified during the ED evaluation were also identified as was the disposition following the evaluation. Laboratory values for potassium, ionized calcium, phosphorus, and magnesium were also reviewed to verify ED diagnoses. All analyses were performed using SAS software, version 9.2 (SAS Institute Inc). Univariate 2-group t tests, Wilcoxon Mann-Whitney tests, and χ² tests were used, as appropriate, based on data type and distribution. The median time between surgery and ED visit was calculated using the product-limit survival method. Multiple logistic regression predicted the odds of an ED visit based on presence of DM, GERD, or PPI use. Odds ratios (ORs) and 95% CIs were considered significant at P < .05.

There were 570 unique patients identified from the thyroid/parathyroid database with a total of 574 surgeries between January 1, 2009, and October 7, 2010. Of these patients, a total of 64 unique patients visited the ED within 30 days of surgery, with a total of 75 visits. This corresponds to an 11% rate of ED evaluation within the first 30 days following thyroid or parathyroid surgery, if multiple visits are not taken into account. One person was evaluated in the ED 3 separate times during this period, and 4 people were evaluated twice.

As summarized in Table 1, the demographic characteristics for patients who visited the ED and those who did not were similar in age distribution, race, BMI, and medical comorbidities such as HTN and ESRD (P > .05 for all comparisons). The presence of medical comorbidities such as DM (P = .03) and GERD (P = .04) and perioperative outpatient use of PPIs (P = .006) were noted to be significantly different between each of the groups.

By multiple logistic regression analysis of the statistically significant factors of DM, GERD, and PPI use, we found that patients taking a PPI were 1.81 times more likely to visit the ED compared with those not taking a PPI (P = .04). Patients with DM had an OR of 1.78 that was not statistically significant (P = .054).

The types of surgical procedures were not significantly different between those patients who required postoperative ED evaluation (Table 2) and those who did not (P = .11). The mean (SD) length of hospital stay following surgery was also not significantly different between the 2 groups, with the no-ED groups staying 1.05 (2.5) days and the ED groups staying 1.25 (2.5) days.

Additional factors that were not statistically different in each group included whether central or lateral neck dissections were completed (P = .32), the number of parathyroid glands excised (P = .53), the finding of malignancy in the pathology specimen (P = .91), and whether parathyroid autotransplant was performed (P = .15).

The median timing of the ED presentation postoperatively was day 5 (95% CI, day 4–day 7). Symptoms triggering the ED evaluation are summarized in Table 3. By far, the most common trigger for an ED evaluation was paresthesia (n = 28). Of patients with paresthesia, most had no electrolyte abnormalities detected of calcium or magnesium (n = 12), while the remainder had hypocalcemia (n = 6), hypomagnesemia (n = 5), or a combination of hypomagnesemia and hypocalcemia (n = 5). Patients with paresthesia presented significantly sooner to the ED than did the rest of the ED patients, as illustrated by the Kaplan-Meier curves in the Figure (P = .003).

As detailed in Table 4, the most common diagnosis given to the group of patients evaluated in the ED was electrolyte abnormalities of calcium, magnesium, phosphorus, or potassium. There were no statistical associations between PPI use and paresthesias (P = .39), confirmed hypokalemia (P = .74), hypomagnesemia (P = .43), or hyperphosphatemia (P = .75). Nor was PPI use a significant factor in the group of patients who presented to the ED with any combination of these electrolyte abnormalities (P = .69).

As summarized in Table 5, most patients were sent home following their ED evaluation and treatment. Several patients required a short stay in the clinical decision unit, while only 3 patients required full inpatient admission.

With the trend toward outpatient thyroidectomy and parathyroidectomy surgery, perioperative complications will be more likely to appear and be managed through the ED. Missing among the literature is a description of the postoperative issues requiring an ED evaluation in a large collection of patients who have undergone thyroid or parathyroid surgery. These issues are important because they affect the overall cost and efficiency of patient care.

In this study, a significant percentage of patients required a visit to the ED during the first 30 days postoperatively. The reasons for the ED evaluations were varied, but a significant number were related to perioperative electrolyte disturbances. With the exception of patients undergoing lobectomy, all patients were started on calcium supplementation postoperatively, but the amount, type, and compliance rates varied widely, and thus this factor was not assessed. Magnesium levels were intermittently tested and rarely supplemented without diagnosed hypomagnesemia. However, there was a fairly large subset of patients who presented to the ED with paresthesias but had normal ionized calcium levels with hypomagnesemia or no detectible abnormality of their serum calcium and magnesium levels.⁹

Patients taking PPIs during the postoperative period were statistically more likely to visit the ED than those not taking this medication. Subgroup analysis of the relationship between those diagnosed as having an electrolyte abnormality and PPI use did not suggest a relationship between the two; however, changes in enteral electrolyte absorption with PPI use has been documented in the literature and is attributed to several factors, although the exact mechanism has not be elucidated.⁸ Changes in gastric acid secretion affects the absorption of many nutrients and vitamins owing to the manipulation of the normally low pH environment.^10-13 It is assumed that long-term PPI use leads to decreased intestinal absorption of calcium owing to lower solubility of calcium salts in the higher gastric pH environment.¹³ Improvement in calcium absorption has been demonstrated in an gastrectomized animal model with administration of dietary lactic acid.¹¹ Unfortunately, in human studies, there have been conflicting results with regard to calcium absorption in the setting of altered gastric pH and PPI use.^13-15 This altered pH environment could also contribute to postoperative ED evaluations owing to altered medication absorption/clearance and changes in the gastric mucosal barrier.^16,17

Other possible mechanisms include PPI-induced hypergastrinemia resulting in parathyroid hyperplasia and increased parathyroid hormone (PTH) secretion. The physiologic feedback loop behind this effect is as follows: by inhibiting gastric acid secretion, somatostatin release is inhibited as well as its negative feedback on gastrin secretion.¹⁵ During the first few months of PPI therapy, serum gastrin levels can increase to 2- to 6-fold that of normal in a majority of patients.¹⁸ Hypergastrinemia has a stimulatory effect on the parathyroid glands, increasing their volume, weight, and PTH hormone gene expression in animal models.^19,20 In a human study, omeprazole administration for 8 weeks caused an increase in PTH level by 27%, but there were no concurrent evaluations of gastrin levels during this study.²¹ This chronic PTH elevation is thought to be one of the mechanisms for increased hip and spine fractures seen with long-term PPI use.²² Given the changes seen in PTH levels with thyroidectomy and parathyroidectomy, it is possible that this hypergastrinemia-induced PTH elevation could have a significant effect on postoperative calcium levels.

A significant relationship was seen between the presence of DM and the need for ED evaluation, but this was not retained when accounting for the presence of GERD or PPI use. Further subgroup analysis was not completed on this group of patients. There is a significant amount of literature describing the relationship of DM with poor surgical outcomes, even in head and neck patients.^23-25 With the higher rates of obesity and increasing prevalence of DM, it is imperative that medical optimization is sought by the surgical team preoperatively.

While most of the patients in our study were managed in the ED and discharged home, as detailed in Table 1, about 20% required admission to at least an observation unit. It is interesting that the median timing for the ED evaluation was around the fifth postoperative day, but when the ED group was separated by the presence or absence of paresthesias, this median timing was separated to the third and eighth postoperative day, respectively. Given the significant percentage of ED visits associated with electrolyte abnormalities, this postoperative period was likely necessary to reach a new equilibrium in the postoperative state.

A drawback of this study is the widely different methods of calcium supplementation used during the postoperative period. While almost all of these patients were instructed to take postoperative calcium, different dosing methods, formulations, and levels of medication compliance were not taken into account. Recommended postoperative calcium replacement regimens are not uniform.^6,7,26,27 The role of postoperative parathyroid hormone testing to determine the appropriate aggressiveness of replacement vs the cost-effectiveness of such testing is still debated.^28-30

While there was no statistically significant association seen between PPI use and the diagnosis of an electrolyte abnormality or paresthesia in the ED, this subgroup of patients was small. Based on the results of this study, it would be prudent to consider delaying initiation of PPI therapy in the early postoperative period; however, the duration of PPI treatment necessary to cause the increased risk of ED evaluation is unknown. It is also not known if this increased risk is seen with different types of reflux therapy, such as H₂ blockers. The effects of PPI therapy in the post-thyroidectomy or post-parathyroidectomy state warrants further investigation.

The possibility of ED evaluations outside of the Henry Ford Health System should also be considered as a limitation of the study. This limitation would mean that the reported rate of ED visits underestimates the true postoperative ED evaluation rate, but this should be minimized because the Henry Ford Health System is a large regional network that include multiple satellite ED locations with a common electronic medical record system.

In conclusion, postoperative complications from thyroidectomy and parathyroidectomy requiring ED evaluation are significant. While the most common cause of ED evaluation within 30 postoperative days was electrolyte abnormalities, postoperative patients taking PPIs were found to be more likely to visit the ED than patients not taking PPIs. Although the specific causes of this increased risk is unknown, changes in the absorption of electrolytes, medications, and even the permeability of the gastric lining itself have been documented with PPI use and may be associated with higher ED evaluation rates following thyroidectomy or parathyroidectomy.

Corresponding Author: Tamer A. Ghanem, MD, PhD, Department of Otolaryngology–Head and Neck Surgery, Henry Ford Hospital, 2799 W Grand Blvd, Detroit, MI 48202 (tghanem1@hfhs.org).

Published Online: September 5, 2013. doi:10.1001/jamaoto.2013.4505.

Author Contributions: Dr Young had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Young, Succar, Ghanem.

Acquisition of data: Young, Succar, Hsu, Talpos.

Analysis and interpretation of data: Young, Talpos, Ghanem.

Drafting of the manuscript: Young, Succar, Hsu, Ghanem.

Critical revision of the manuscript for important intellectual content: Talpos, Ghanem.

Statistical analysis: Young, Succar.

Administrative, technical, or material support: Young.

Study supervision: Talpos, Ghanem.

Conflict of Interest Disclosures: None reported.

Previous Presentation: This study was presented at the American Head and Neck Society 2013 Annual Meeting; April 11, 2013; Orlando, Florida.

Additional Contributions: Meredith Mahan, MS, Department of Public Health Sciences, Henry Ford Health System, Detroit, Michigan, assisted with statistical analysis. Ms Mahan is paid for her position by her department. In addition, there exists interdepartmental funding for the study statistics.