Abstract
Objectives:
To investigate the relationship between thyroid function and short-term outcomes in elderly patients undergoing surgery for hip fracture.
Methods:
Patients aged >60 years admitted from July 2009 to June 2010 who underwent surgical fixation of low-energy hip fracture, excluding those with pathological or periprosthetic fracture, were enrolled in a retrospective observational cohort study. Patients were classified as having biochemically overt or subclinical hyperthyroidism or hypothyroidism, normal thyroid function, or nonclassifiable state based on preoperative thyroid-stimulating hormone and free thyroxine values. Outcome data were collected from hospital records. Biochemical thyroid dysfunction was not systematically treated. Outcomes measured were length of stay (LOS), 30-day postoperative complications, readmissions, and mortality.
Results:
A total of 254 patients were analyzed; 64 (25.2%) were male and mean age was 77.8 years. There were 128 (50.4%) patients with euthyroid, 68 (26.8%) nonclassifiable, 13 (5.1%) overtly hyperthyroid, 20 (7.9%) subclinically hyperthyroid, 4 (1.6%) overtly hypothyroid, and 21 (8.3%) subclinically hypothyroid. The 30-day postoperative complication rate was 38% (96 patients). The most common complication was urinary tract infection (12.6%), followed by cardiac events (8.3%) and delirium (5.5%). Patients with overt hyperthyroidism had an increased risk of complications in multivariate analysis (odds ratio 3.75, 95% confidence interval 1.10-12.84). Complications in this group were similar to those in the overall cohort. Thyroid function did not predict LOS, readmissions, or mortality.
Conclusion:
Older patients frequently develop complications following surgery for hip fractures. This risk appears to be increased by preoperative biochemically overt hyperthyroidism. Further study is warranted to confirm this finding and to determine whether preoperative treatment improves outcomes.
Introduction
Osteoporotic hip fractures in the elderly individuals are a major clinical burden in Asia. 1 Patients are at high risk of adverse events, both in the immediate perioperative period and in the long term. Surgical fixation of hip fractures in the elderly individuals constitutes a major physical stressor, frequently exacerbated by a reduced physiological reserve and the presence of medical comorbidities.
Some of the known risk factors for morbidity and mortality include the presence and number of medical comorbidities, age, body mass index as well as several biomarkers such as cardiac troponin I, serum albumin, tumor necrosis factor-α, interleukin 6 and 10, and intact parathyroid hormone (iPTH). 2 –9 Derangements in thyroid function tests in the elderly are common; however, the existing literature is conflicting as to their clinical significance. In our literature review, we found only 1 article investigating the impact of subclinical thyroid dysfunction on outcomes after hip fracture in the elderly patients which showed no difference in mortality at 1 year. 10 However, this study did not capture other secondary outcomes such as complication rate and may have been underpowered to detect small differences in mortality. We postulated that patients with thyroid dysfunction might be at increased risk of complications, in particular cardiac-related events, leading possibly to an increased length of stay and early (30-day postoperative) mortality. Therefore, we set out to design a study to explore this clinical question.
Methods
We conducted a retrospective observational cohort study involving patients admitted to the orthopedic unit of Singapore General Hospital from July 2009 to June 2010 with a diagnosis of low-energy hip fracture. Excluded were patients aged <60 years, those with pathological or periprosthetic fracture, and those whose fractures were not treated operatively.
Serum thyroid-stimulating hormone (TSH) and free thyroxine (fT4) along with baseline renal function, serum iPTH, albumin, and calcium were measured on admission as part of a coordinated clinical pathway for patients with hip fracture. Outcome data, patient demographics, and past medical history were collected from hospital records and discharge summaries. We examined specific outcome measures, including length of stay (LOS) in hospital, 30-day postoperative medical and surgical-related complications, readmission rate, and mortality. Serum TSH and fT4 were measured using the immunoassay method with an interassay coefficient of variability of 5.4% and 3.1%, respectively (Beckman Coulter, Inc, Brea, California). The normal reference range that we used was 0.65 to 3.70 µIU/mL for TSH and 0.68 to 1.12 ng/dL for fT4. Patients were classified into 6 categories based on their biochemical thyroid profile as illustrated in Table 1. We were not able to correlate this with clinical manifestations of thyroid hypo- or hyperactivity, as this information was not collected uniformly. Patients with thyroid function test abnormalities were also not systematically treated or referred for specialist endocrinology opinion.
Thyroid Status Classification Based on Thyroid Function Tests.
Abbreviations: fT4, free thyroxine; TSH, thyroid-stimulating hormone; LLN, lower limit of normal, ULN, upper limit of normal.
The presence of end-stage renal failure was defined as either a calculated glomerular filtration rate of <15 mL/min by the Cockcroft-Gault equation or the requirement for chronic hemodialysis.
Statistical analysis was done using MedCalc for Windows, version 12.1.4.0 (MedCalc Software, Mariakerke, Belgium). The student's t test and the Mann-Whitney test were used to compare the means of normally and non-normally distributed parameters, respectively. The association with age, gender, preexisting comorbidities, thyroid status, serum albumin, serum vitamin D, serum iPTH level, and end-stage renal failure with regard to 30-day complications, readmissions, and mortality was analyzed using logistic regression. The association with the above-mentioned factors and length of hospital stay was analyzed using multiple regression. The P value cutoff for significance was .05.
A total of 254 patients were included in the final analysis. The baseline characteristics of the cohort are detailed in Table 2. The mean age for males and females was 76.0 and 78.5 years, respectively. This difference was not statistically significant (P =.052) as determined by the independent samples t test. The female to male ratio was approximately 3:1. The majority (83.5%) of patients were Chinese, representative of our local demographics. The distribution between neck of femur and intertrochanteric fractures was equal. More than half (55.1%) of the cohort had at least 1 significant medical comorbidity, the commonest being diabetes mellitus. There were 9 patients with a preexisting known thyroid disorder, of which 8 had hypothyroidism and 1 had hyperthyroidism.
Baseline Preoperative Characteristics.
Abbreviation: iPTH, intact parathyroid hormone.
Results
Based on the admission thyroid function tests, 50.4% of the patients were biochemically euthyroid, 8.3% were subclinically hypothyroid, 1.6% were overtly hypothyroid, 5.1% were subclinically hyperthyroid, and 7.9% were overtly hyperthyroid; 26.8% of the patients were nonclassifiable according to our proposed categories of subdivisions. The majority of patients in the nonclassifiable group displayed high fT4 and normal TSH values as illustrated in Figure 1. The absolute fT4 and TSH values by category are presented in Table 3.
Scatter plot of all patients of free T4 against thyroid-stimulating harmone (TSH). X-axis: free T4 (ng/dL), Y-axis: TSH (µIU/mL). Dotted lines represent upper and lower limits of normal for TSH and free T4, respectively. The shaded area denotes the range of values of TSH and free T4 that fall under the nonclassifiable category.
Distribution of Patients by Thyroid Status.a
Abbreviations: fT4, free thyroxine; TSH, thyroid stimulating hormone.
a All values in µIU/mL for TSH and ng/dL for fT4. “Low” is defined as being below the lower limit of reference range and “high” is defined as being above the upper limit of reference range.
b Overt hyperthyroidism: low TSH, high fT4, subclinical hyperthyroidism: low TSH, normal fT4, euthyroid: normal TSH and fT4, subclinical hypothyroidism: high TSH, normal fT4, overt hyperthyroidism: high TSH, low fT4, nonclassifiable: all other combinations of TSH and fT4.
Of the 254 patients who were studied, 96 (38%) had at least 1 complication in the 30 days following surgery, of which the top 3 complications were urinary tract infection, composite cardiovascular event (including congestive cardiac failure, acute coronary syndrome, and other events including arrhythmias), and delirium as detailed in Table 4. The median LOS on the orthopedic ward was 11 days. Readmissions and mortality at 30 days were 9.6% and 1.2%, respectively.
Distribution of 30-Day Postoperative Complications.
Univariate and multivariate analyses of the associations of age, gender, fracture type, preexisting comorbidities, thyroid status, serum vitamin D and parathyroid hormone level, serum albumin, and renal function with respect to 30-day complications, LOS, and 30-day mortality were performed. The findings are presented in Tables 5 to 7.
Univariate Factor Analysis of Variables for 30-Day Complications, Readmissions, and Length of Stay.a
Abbreviations: CI, confidence interval; iPTH, intact parathyroid hormone; LOS, length of stay; NA, not applicable; OR, odds ratio.
a Statistically significant factors are in bold.
b Complications were defined as any adverse event resulting from the fracture, surgery, or hospitalization. Refer to Table 4 for the full list of complications observed.
c ORs for the above values were not computed, as no events were observed in that category.
Logistic Regression Analysis of Variables for 30-Day Complications and Readmissions.a
Abbreviations: CI, confidence interval; iPTH, intact parathyroid hormone; NA, not applicable; OR, odds ratio.
a Statistically significant factors are in bold.
b ORs for the above were not computed as no events were observed in that category.
Multiple Regression Analysis of Variables for LOS. Statistically Significant Factors are in Bold.
Abbreviations: CI, confidence interval; iPTH, intact parathyroid hormone; LOS, length of stay; NA, not applicable.
With respect to 30-day complications, defined as the presence of at least 1 complication, biochemical overt hyperthyroidism was associated with an increased risk (odds ratio (OR) 3.75, 95% confidence interval 1.10-12.84) relative to the euthyroid group on multivariate analysis. The distribution of the type of complications seen in the overtly hyperthyroid group did not differ significantly from the euthyroid cohort. None of the ORs for the other thyroid categories reached statistical significance. Other significant risk factors were male sex, end-stage renal failure, and increasing age.
Thyroid status on admission did not significantly affect LOS or 30-day complications. The factors that were associated with increased LOS were increasing age, intertrochanteric fracture, and end-stage renal failure, while only age was significantly associated with a higher risk of readmission at 30 days.
Due to the low incidence of mortality at 30 days, none of the factors reached statistical significance given the sample size studied.
Discussion
A significant proportion of hospitalized patients with nonthyroidal illness (NTI) present with derangements in thyroid function tests. 11,12 It has been postulated that these changes in thyroid function during severe illness are protective in that they prevent excessive tissue catabolism. 13,14 The possible underlying reasons for these abnormalities in thyroid function include inhibition of 5’-monodeiodination of T4 to T3, 15 reductions in the serum concentrations of thyroid hormone-binding proteins, 16 an increase in circulating substances that inhibit T4 binding to binding proteins, 17 and acquired transient central hypothyroidism, 13 all of which may result in a myriad of possible combinations of thyroid function abnormalities. Although a small proportion of the patients in this cohort had a known preexisting thyroid disorder, we cannot preclude the possibility of undiagnosed clinical hyper or hypothyroidism, especially since the manifestations tend to be more muted in the elderly population. 18 However, it is equally, if not more likely, that the acute stress of a hip fracture would cause changes in their fT4 and TSH levels through the effects of NTI. Therefore, the major limitation of our study is that the categorization of patients based on their admission thyroid function tests may not have accurately grouped them according to their “true” thyroid status.
Moreover, the majority of patients in the “nonclassifiable” group did not demonstrate the typical pattern found in NTI of low fT4 and TSH, but were instead found to have high fT4 and normal or high TSH as demonstrated in Figure 1. In our cohort, undiagnosed secondary hyperthyroidism as a cause of high fT4 and TSH is a possibility; however, we did not further evaluate these cases due to the retrospective nature of the study and the fact that they did not present with signs and symptoms suggestive of pituitary disease. In addition, TSH may be elevated transiently during the recovery phase of NTI, 19 and fT4 may also be falsely high or low in NTI depending on the assay used. 20
Nevertheless, our study demonstrates that biochemical overt hyperthyroidism present on admission portends a higher risk (4-fold) of at least 1 medical or surgical complication within the first 30 days after undergoing hip fracture surgery. This can potentially serve as a red flag warranting increased vigilance among the managing physicians and surgeons during the early postoperative period. This was however not associated with an increase in LOS. We also did not note any increase in the 30-day mortality in those with biochemical hyperthyroidism (or other categories of dysfunction, as our study was likely to be underpowered to examine the relationship of thyroid dysfunction with mortality. We note that the types of complications seen did not appear significantly different from the euthyroid group. Therefore, there did not seem to be any predilection for cardiac-related complications, which might have been expected in a situation of thyroid hormone excess. However, thyroid hormones affect many different tissues accounting for their widespread systemic manifestations. For instance, thyroid hormones act as modulators of the immune response and both hypothyroidism and hyperthyroidism can alter immune functions such as chemotaxis, phagocytosis, generation of reactive oxygen species, and cytokine synthesis and release. 21
Our study raises an interesting novel observation linking biochemical hyperthyroidism with early 30-day complications in patients with hip fracture. A larger, prospective study with a protocolized pathway for further evaluation of patients with abnormal thyroid function tests, including measurement of admission T3 levels, clinical correlation of thyroid status as well as subsequent follow-up with repeat thyroid function test after the acute stress, might help to clarify our preliminary findings.
Footnotes
Acknowledgments
This study was previously presented in poster format at ENDO 2012: The Endocrine Society’s 94th Annual Meeting & Expo.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.