Impact of Hypothyroidism on Short-Term and Long-Term Outcomes After Lumbar Fusion: A Nationwide Propensity-Matched Cohort Study

Abstract

Study Design

Retrospective cohort study.

Objectives

Hypothyroidism is associated with increased perioperative risks across various surgical specialties and has recently been linked to hematologic and short-term complications following spinal fusion. However, its impact on long-term mechanical complications remains unclear. This study aims to evaluate short-term (45-day) and long-term (2-year) medical and mechanical complications in hypothyroid patients undergoing primary lumbar fusion.

Methods

A retrospective analysis was conducted using the TriNetX database to identify patients who underwent lumbar fusion between 2002 and 2022 with at least 2 years of follow-up. Patients with preoperative hypothyroidism (n = 3,348, identified via ICD-10-CM code E03.9) were propensity score–matched 1:1 to euthyroid controls based on demographic and clinical risk factors.

Results

After matching (n = 2850 per cohort), hypothyroid patients had higher rates of sepsis at 45 days (2.1% vs 1.3%; P = 0.019; RR: 1.62, 95% CI [1.08-2.44]). At 2 years, they had increased risks of chronic device-related infection (0.7% vs 0.3%; P = 0.019; RR: 2.27, 95% CI [1.12-4.61]), wound disruption (3.6% vs 2.7%; P = 0.031; OR: 1.35, 95% CI [1.03-1.75]), and proximal junctional kyphosis (3.2% vs 2.3%; P = 0.017; RR: 1.42, 95% CI [1.06-1.90]). No significant differences were observed in readmissions, revision surgery, or pseudoarthrosis rates.

Conclusions

Hypothyroidism is a risk factor for early and late postoperative complications in patients undergoing primary lumbar fusion. Appropriate preoperative optimization with thyroid hormone supplementation to achieve euthyroid status could potentially minimize the incidence of developing notable postoperative medical and mechanical complications.

Keywords

hypothyroidism lumbar fusion postoperative complications sepsis spine infection proximal junctional kyphosis

Background

Spinal fusion, initially developed to address instability caused by tuberculosis or spinal deformities, has evolved significantly over time. With advancements in surgical techniques, its indications have expanded to treat a broad spectrum of conditions, including traumatic injuries, deformities, primary and secondary tumors, infections, and degenerative spinal disorders. As a result, the number of lumbar fusions performed worldwide has increased rapidly in recent years.^1,2 Despite its widespread use in managing complex lumbar spine pathologies, postoperative complications remain a significant concern, particularly in patients with comorbidities that may affect recovery.³

Thyroid hormones play an important role in the maturation of the skeleton, as the skeleton is profoundly sensitive to thyroid hormones for bone development, growth, and overall structural maintenance.⁴ Hypothyroidism, characterized by thyroid hormone deficiency, is an endocrinopathy that continues to increase in prevalence affecting many patients worldwide, with a prevalence as high as 20.6% in certain populations.⁵ Previous studies have demonstrated an increased risk of fracture in patients with hypothyroidism, secondary to the pathophysiological change of bone turnover due to alteration in thyroid hormone levels.^6,7 Hypothyroidism has also been shown to increase the risk of surgical complications, such as delayed wound healing.^8,9 Research within other surgical specialties such as primary total knee arthroplasty has demonstrated the role of hypothyroidism in increasing the risk of periprosthetic infections, and multiple postoperative complications notably in the perioperative and 90-day postoperative period.¹⁰ Studies evaluating spinal fusion procedures suggest that hypothyroidism may contribute to elevated 90-day complication rates consisting of wound complications and pulmonary emboli, particularly in patients undergoing primary anterior cervical discectomy and fusion (ACDF) surgery.⁹

Despite the short-term and long-term impacts of hypothyroidism that have been studied in other surgical subspecialties, the impact of hypothyroidism on postoperative outcomes following primary lumbar fusion surgeries remains scarce in the current literature. Short-term postoperative medical outcomes of lumbar fusion surgery in patients with hypothyroidism have been studied in comparison to euthyroid counterparts but did not examine long-term medical or mechanical outcomes in this patient population.¹¹ Mechanical complications such as implant failure, proximal junctional kyphosis, and pseudoarthrosis typically manifest beyond the immediate postoperative period.¹² Hence, there remains a significant gap in the current literature with regards to early medical and long-term postoperative mechanical complications in hypothyroid patients undergoing primary lumbar fusion. This study specifically aims to address this gap by examining outcomes up to 2 years.

This study aimed to determine whether patients with hypothyroidism are at greater odds of developing 45-day medical complications (early) and 2-year mechanical complications (late) compared to patients without hypothyroidism. We hypothesize that patients with clinical hypothyroidism are at greater odds of developing early postoperative medical and late mechanical complications than those without hypothyroidism. By identifying these potential risks, our findings may help refine perioperative screening and management for hypothyroid patients undergoing LF, ultimately informing future surgical risk stratification and postoperative monitoring guidelines.

Material and Methods

Study Design

This retrospective cohort study examined the association between hypothyroidism and outcomes after lumbar fusion (LF) using de-identified electronic medical records. Data queries were executed on January 27, 2025, to identify patients undergoing LF within a 20-year period prior to December 31, 2022, and assess outcomes. As this study relied solely on de-identified patient records without individually identifiable data, it was deemed exempt from Institutional Review Board approval.

Setting and Data Source

Data were sourced from TriNetX, a US-based federated health research network comprising de-identified electronic health records (EHRs) from over 100 million patients across more than 100 healthcare organizations. TriNetX aggregates and standardizes data from participating HCOs using controlled terminologies (eg, ICD, CPT, LOINC, RxNorm), enabling researchers to perform queries across the network. It is important to note that data quality, completeness, and granularity depend on the documentation practices within the originating HCOs, and specific institutional data verification protocols are generally not publicly available. Data queries were performed on January 27, 2025. While specific granular details regarding data verification processes or data quality assurance protocols at each contributing institution are not available, TriNetX employs data curation and standardization methodologies to enhance consistency across the federated network.

Participants

Adults (aged 18 to 90 years) who underwent lumbar fusion procedures between 2002 and 2022 were identified using CPT codes and were subsequently divided into cohorts according to the presence or absence of hypothyroidism. Patients were excluded if they had records of lumbar fracture, fracture fixation, vertebral neoplasm, vertebral corpectomy, lumbar laminectomy, or hyperthyroidism.

Eligibility Criteria

Patients were included if they underwent lumbar fusion procedures identified by specific CPT codes: anterior interbody fusion (22 558), posterior interbody fusion (22 630, 22 633), posterior or posterolateral fusion (22 612), and lateral extracavitary fusion (22 533). Patients in the hypothyroidism cohort needed to have a diagnosis of hypothyroidism (ICD-10-CM: E03.9), documented between 1 year and 1 month prior to undergoing lumbar fusion. We required patients in the control cohort to have no prior diagnosis of hypothyroidism (ICD-10-CM: E03.9) or hyperthyroidism (ICD-10-CM: E05).

Cohorts

This study examined 2 cohorts based on the presence or absence of hypothyroidism. Prior to propensity matching, the hypothyroidism cohort (n = 3363) and the control cohort (n = 37 289) exhibited significant differences in baseline demographics and comorbidities.

Variables

To minimize bias, we used propensity score matching (1:1) to control for confounding variables associated with outcomes after lumbar fusion. Matched variables included demographics (ie, age at index surgery, sex, race, ethnicity), BMI, tobacco use, and comorbidities including chronic kidney disease, diabetes mellitus, hypertensive diseases, heart failure, liver disease, overweight/obesity, and inflammatory polyarthropathies. After propensity score matching, each cohort consisted of 2850 patients with balanced baseline characteristics (all standardized differences < 0.1).

Outcomes

Our primary outcomes were categorized as short-term medical and mechanical complications (within 45 days postoperatively) and long-term mechanical outcomes (within 2 years postoperatively). Short-term medical complications included sepsis, infection, pulmonary embolism (PE), deep vein thrombosis (DVT), myocardial infarction (MI), mortality, hospital readmission, and renal failure. Long-term mechanical complications included reoperation, chronic infection due to device, spine infection comprising of discitis, osteomyelitis and intraspinal abscess, pseudoarthrosis, proximal junctional kyphosis (PJK), and proximal junctional failure (PJF). Outcomes were ascertained by querying for relevant ICD-10-CM, ICD-9-CM, and CPT codes within the respective postoperative timeframes. ICD codes for each postoperative outcome can be found in Table 1.

Table 1.

ICD Codes for Postoperative Outcomes.

Medical Complication Codes
Outcome	Codes
Sepsis	A41.9; T81.44
Infection	T81.4; T81.43XA
PE	I26
DVT	I82.40
MI	I21
Mortality	Demographics: Deceased
Hospital readmission	99 221; 99 222; 99 223
Acute renal failure	N17; N19; N99.89
Mechanical complication codes
Outcome	Codes
Reoperations	22 612; 22 558; 22 630; 22 633; 22 533
Chronic infection due to device	T84.7XXD
Seroma	L76.34; M96.842; G97.63; G97.64
Hematoma	L76.32; M96.840; G97.61; G97.62
Wound disruption	T81.31XA
Spine infection	M46.4; M46.2; G06.1
Pseudoarthrosis	M96.0
Proximal junctional kyphosis	M40.00; M40.2
Proximal junctional failure	M96.89; T84.89XA; M48.8 × 9
Postlaminectomy syndrome	722.83
CSF Leak	G96.0; G96.1

Statistical Analysis

Initial statistical analysis was completed using the built-in statistical functions of the TriNetX database software. For comparisons of patient baseline demographics, this study used independent-samples t-tests for continuous variables and Pearson chi-squared tests for categorical variables. We used standardized mean difference (SMD) to evaluate meaningful between cohort differences with a threshold of SMD < 0.1 indicating balance. The Measure of Association analysis package within TriNetX was utilized to compare the proportion of patients experiencing each outcome between the hypothyroidism and control cohorts post-matching. This generated risk ratios (RRs) and risk differences (RDs) with 95% confidence intervals (CIs). Statistical significance for the risk difference was based on the z-statistic and corresponding P-value, with P < 0.05 considered significant. For reoperation outcomes, Kaplan-Meier survival analysis and Log-Rank test were additionally performed to assess time-to-event data.

Results

Patients

We identified 40 652 eligible patients from the TriNetX Research network. Before propensity matching there were 3363 patients in the hypothyroidism cohort and 37 289 in the control cohort. As noted in Table 2, patients in the hypothyroidism cohort before matching were significantly older (mean age 64.3 vs 57.9 years, P < .001) and had a markedly higher prevalence of matched comorbidities, including hypertension (53.1% vs 21.6%, P < .001), diabetes mellitus (23.2% vs 9.0%, P < .001), chronic kidney disease (8.9% vs 2.8%, P < .001), and obesity (21.0% vs 8.3%, P < .001). This highlights the strong baseline association between hypothyroidism and factors related to metabolic syndrome within this database population. Following matching, each cohort had 2850 patients. Key variables were optimally balanced (SMD < 0.1 for most variables) with no significant differences in mean age (64.3 vs 64.2 years, P = .785), sex, race, ethnicity, or the matched comorbidities (Table 2).

Table 2.

Patient Characteristics Propensity Matching.

Patient Characteristics Before and After Propensity Score Matching
Demographics & comorbidities	Cohort	# Patients (Before)	% Cohort (Before)	P-value (Before)	# Patients (After)	% Cohort (After)	P-value (After)
Age at index	Hypothyroidism & LF	2866	100%	<0.001	2850	100%	0.785
Age at index	Control	36 175	100%	<0.001	2850	100%	0.785
Gender
Female	Hypothyroidism & LF	2125	74.60%	<0.001	2125	74.60%	0.649
Female	Control	17 469	49.40%	<0.001	2110	74.00%	0.649
Male	Hypothyroidism & LF	602	21.10%	<0.001	602	21.10%	0.605
Male	Control	16 800	47.50%	<0.001	618	21.70%	0.605
Race/Ethnicity
White	Hypothyroidism & LF	2335	81.90%	<0.001	2335	81.90%	0.184
White	Control	25 344	71.60%	<0.001	2373	83.30%	0.184
Black or African American	Hypothyroidism & LF	177	6.20%	<0.001	177	6.20%	0.912
Black or African American	Control	3807	10.80%	<0.001	175	6.10%	0.912
Asian	Hypothyroidism & LF	44	1.50%	0.069	44	1.50%	0.433
Asian	Control	722	2.00%	0.069	37	1.30%	0.433
American Indian/Alaska Native	Hypothyroidism & LF	10	0.40%	0.269	10	0.40%	1
American Indian/Alaska Native	Control	86	0.20%	0.269	10	0.40%	1
Unknown race	Hypothyroidism & LF	224	7.90%	<0.001	224	7.90%	0.584
Unknown race	Control	4070	11.50%	<0.001	213	7.50%	0.584
Ethnicity
Not Hispanic or Latino	Hypothyroidism & LF	2175	76.30%	<0.001	2175	76.30%	0.186
Not Hispanic or Latino	Control	25 359	71.70%	<0.001	2217	77.80%	0.186
Hispanic or Latino	Hypothyroidism & LF	137	4.80%	0.002	137	4.80%	0.41
Hispanic or Latino	Control	2213	6.30%	0.002	124	4.40%	0.41
Comorbidities
Body mass index [BMI]	Hypothyroidism & LF	598	21.00%	<0.001	598	21.00%	0.695
Body mass index [BMI]	Control	3186	9.00%	<0.001	586	20.60%	0.695
Tobacco use	Hypothyroidism & LF	61	2.10%	<0.001	61	2.10%	0.34
Tobacco use	Control	463	1.30%	<0.001	51	1.80%	0.34
Chronic kidney disease	Hypothyroidism & LF	255	8.90%	<0.001	255	8.90%	0.084
Chronic kidney disease	Control	1006	2.80%	<0.001	219	7.70%	0.084
Diabetes mellitus	Hypothyroidism & LF	660	23.20%	<0.001	660	23.20%	0.342
Diabetes mellitus	Control	3201	9.00%	<0.001	630	22.10%	0.342
Hypertensive diseases	Hypothyroidism & LF	1513	53.10%	<0.001	1513	53.10%	0.276
Hypertensive diseases	Control	7643	21.60%	<0.001	1554	54.50%	0.276
Heart failure	Hypothyroidism & LF	139	4.90%	<0.001	139	4.90%	0.026
Heart failure	Control	538	1.50%	<0.001	105	3.70%	0.026
Liver disease, unspecified	Hypothyroidism & LF	19	0.70%	<0.001	19	0.70%	0.143
Liver disease, unspecified	Control	100	0.30%	<0.001	11	0.40%	0.143
Overweight/obesity	Hypothyroidism & LF	599	21.00%	<0.001	599	21.00%	0.513
Overweight/obesity	Control	2929	8.30%	<0.001	579	20.30%	0.513

Descriptive Data

The demographic characteristics were well-balanced between the matched cohorts. The proportion of females was 74.6% in the hypothyroidism cohort and 74.0% in the control cohort (P = .649). The mean follow-up time was 538 days for the hypothyroidism group and 517 days for the control group, with corresponding median follow-up times of 730 for both groups. While no minimum follow-up time was set for inclusion, cohorts were matched on follow-up duration, as demonstrated in Supplemental Figure 1. Race distribution showed no significant differences with White patients comprising 81.9% and 83.3% in the hypothyroidism and control cohorts, respectively (P = .131). The prevalence of comorbidities was similar between cohorts: diabetes mellitus (23.2% vs 23.0%, P = .839), hypertensive diseases (53.1% vs 52.8%, P = .842), chronic kidney disease (8.9% vs 8.7%, P = .792), and obesity (21.0% vs 20.8%, P = .851). Post-matching propensity score densities overlapped when graphed, indicating adequate covariate balance for most variables, though inflammatory polyarthropathies showed a marginally higher standardized mean difference (SMD = 0.112) compared to other variables (Table 2).

45-Day Outcomes

Within the 45-day postoperative period, the hypothyroidism cohort had a significantly higher risk of sepsis compared to the control cohort (2.1% vs 1.3%; RR: 1.62, 95% CI: 1.08-2.44; P = .019). There were no statistically significant differences between the cohorts for spine infection comprising of discitis, osteomyelitis, and intraspinal abscess (2.9% vs 2.6%; RR: 1.15, 95% CI: 0.85-1.57; P = .373), pulmonary embolism (1.3% vs 1.2%; RR: 1.09, 95% CI: 0.69-1.71; P = .724), deep vein thrombosis (0.9% vs 0.9%; RR: 0.96, 95% CI: 0.56-1.65; P = .890), myocardial infarction (1.0% vs 0.9%; RR: 1.12, 95% CI: 0.66-1.89; P = .684), mortality (0.7% vs 0.7%; RR: 1.00, 95% CI: 0.54-1.86; P = 1.000), hospital readmission (20.6% vs 19.3%; RR: 1.07, 95% CI: 0.96-1.18; P = .220), and renal failure (4.0% vs 3.3%; RR: 1.22, 95% CI: 0.94-1.60; P = .139). Early reoperations showed no significant difference between groups (3.1% vs 2.7%; RR: 1.13, 95% CI: 0.84-1.52; P = .431). Tables 3 and 4 present detailed information on 45-day medical and mechanical outcomes, respectively.

Table 3.

45-Day Medical Complications.

Medical Complications
Complication	Hypothyroidism & LF (%/# pts)	Control (%/# pts)	Risk Ratio (RR)	95% CI	Risk Difference	P-Value
Sepsis	2.1% (60)	1.3% (37)	1.622	(1.080, 2.435)	0.008	0.019
Infection	2.9% (84)	2.6% (73)	1.151	(0.845, 1.568)	0.004	0.373
PE	1.3% (38)	1.2% (35)	1.086	(0.688, 1.713)	0.001	0.724
DVT	0.9% (26)	0.9% (27)	0.963	(0.563, 1.646)	0	0.89
MI	1.0% (29)	0.9% (26)	1.115	(0.659, 1.889)	0.001	0.684
Mortality	0.7% (20)	0.7% (20)	1	(0.539, 1.855)	0	1
Hospital readmission	20.6% (587)	19.3% (550)	1.067	(0.962, 1.184)	0.013	0.22
Renal failure	4.0% (115)	3.3% (94)	1.223	(0.936, 1.599)	0.007	0.139

Table 4.

45-Day Mechanical Complications.

Mechanical Complications
Complication	Hypothyroidism & LF (%/# pts)	Control (%/# pts)	Risk Ratio (RR)	95% CI	Risk Difference	P-Value
Reoperations	3.1% (88)	2.7% (78)	1.128	(0.836, 1.523)	0.004	0.431
Chronic infection due to device	0.4% (10)	0.4% (10)	1	(0.417, 2.399)	0	1
Seroma	1.0% (29)	1.1% (30)	0.967	(0.582, 1.606)	0	0.896
Hematoma	0.4% (10)	0.4% (10)	1	(0.417, 2.399)	0	1
Wound disruption	2.4% (69)	1.9% (54)	1.278	(0.898, 1.817)	0.005	0.172
Spine infection	2.7% (78)	2.4% (69)	1.13	(0.821, 1.556)	0.003	0.452
Pseudoarthrosis	11.3% (321)	11.4% (325)	0.988	(0.854, 1.142)	−0.001	0.867
Proximal junctional kyphosis	1.3% (36)	0.9% (25)	1.44	(0.867, 2.392)	0.004	0.157
Proximal junctional failure	0.4% (10)	0.4% (10)	1	(0.417, 2.399)	0	1
Postlaminectomy syndrome	0.5% (14)	0.6% (16)	0.875	(0.428, 1.789)	−0.001	0.714
CSF Leak	0.6% (17)	0.7% (19)	0.895	(0.466, 1.718)	−0.001	0.738
Kaplan-Meier survival analysis
Outcome	Group	Patients with outcome	Survival probability at end of time window	Log-rank test (P-value)	Hazard ratio (95% CI)	HR P-value
Reoperations	Hypothyroidism & LF	88	96.89%	0.456	1.123 (0.828, 1.523)	0.361
Reoperations	Control	78	97.23%

2-Year Outcomes

At the 2-year follow-up, the hypothyroidism cohort exhibited a significantly higher risk of chronic infection due to device (0.7% vs 0.3%; RR: 2.27, 95% CI: 1.12-4.61; P = .019), surgical wound disruption (3.6% vs 2.7%; OR: 1.35, 95% CI: 1.03-1.78; P = .031), and proximal junctional kyphosis (PJK) (diagnosed via ICD code) (3.2% vs 2.3%; RR: 1.42, 95% CI: 1.06-1.90; P = .017). There were no statistically significant differences between the cohorts for spine infections comprising of discitis, osteomyelitis, and intraspinal abscess (4.3% vs 3.9%; RR: 1.12, 95% CI: 0.89-1.41; P = .354, pseudoarthrosis (diagnosed via ICD code) (14.5% vs 14.3%; RR: 1.01, 95% CI: 0.90-1.13; P = .889, or proximal junctional failure (PJF) (diagnosed via ICD code) (0.8% vs 0.7%; RR: 1.27, 95% CI: 0.73-2.22; P = .394. For the reoperation outcome, Kaplan-Meier survival analysis showed no statistically significant difference between the hypothyroidism and control cohorts (Log-Rank χ² = 0.228, df = 1, P = .633; HR: 0.95, 95% CI: 0.76-1.18. The 2-year incidence of reoperation was reported as 4.7% in the hypothyroidism cohort and 4.8% in the control cohort at the endpoint in Table 5 (P = .774). Table 5 provides details on 2-year mechanical outcomes.

Table 5.

2-Year Mechanical Outcomes.

Mechanical Complications
Complication	Hypothyroidism & LF (%/# pts)	Control (%/# pts)	Risk Ratio (RR)	95% CI	Risk Difference	P-Value
Reoperations	4.7% (156)	4.8% (161)	0.969	(0.782, 1.201)	−0.001	0.774
Chronic infection due to device	0.7% (25)	0.3% (11)	2.273	(1.120, 4.611)	0.004	0.019
Wound disruption	3.6% (122)	2.7% (91)	1.341	(1.026, 1.751)	0.009	0.031
Spine infection	4.3% (144)	3.9% (129)	1.116	(0.885, 1.409)	0.004	0.354
Pseudoarthrosis	14.5% (484)	14.3% (480)	1.008	(0.897, 1.133)	0.001	0.889
Proximal junctional kyphosis	3.2% (108)	2.3% (76)	1.421	(1.064, 1.898)	0.01	0.017
Proximal junctional failure	0.8% (28)	0.7% (22)	1.273	(0.730, 2.220)	0.002	0.394
Kaplan-Meier survival analysis
Outcome	Group	Patients with outcome	Survival probability at end of time window	Log-rank test (P-value)	Hazard ratio (95% CI)	HR P-value
Reoperations	Hypothyroidism & LF	156	94.61%	0.633	0.948 (0.761, 1.181)	0.76
Reoperations	Control	161	94.29%

Discussion

Lumbar fusion (LF) is one of the most commonly performed surgical procedures to treat various spinal pathologies, such as spondylolisthesis, spinal stenosis, and degenerative disc disease.¹³ However, postoperative complications remain a significant concern among patients with comorbidities.³ Hypothyroidism, a common endocrine disorder, has been linked to impaired skeletal development and an increased risk of infection raising concern about its impact on both short-term and long-term outcomes in lumbar fusion.⁴ Although some studies have examined the relationship between hypothyroidism and cervical and lumbar spinal fusion outcomes, its impact on long-term outcomes in lumbar fusion remain sparse.¹¹ This study examined these gaps by evaluating perioperative (45-day) medical complications, including sepsis, as well as long-term (2-year) mechanical outcomes such as proximal junctional kyphosis (PJK) and chronic device-related infections. Additionally, with the use of propensity score matching to help mitigate numerous established comorbidities in patients with hypothyroidism, our findings build on prior research while also providing a more comprehensive understanding of hypothyroidism’s influence on LF outcomes.

The increased susceptibility of hypothyroid patients to perioperative complications is supported by prior studies. For instance, Vakharia et al analyzed the outcomes of hypothyroid patients undergoing 1-2 level lumbar fusion (1-2 LF) compared to matched controls. Their findings revealed significantly higher rates of 90-day postoperative complications among hypothyroid patients, including infections, deep vein thrombosis (DVT), pulmonary embolism (PE), and pneumonia.¹⁴ Similarly, our study identified an elevated risk of sepsis within the 45-day perioperative period, as well as long-term complications, such as chronic device-related infections, wound disruption, and proximal junctional kyphosis, further reinforcing the heightened vulnerability of this patient population to increased postoperative complications.

The association between hypothyroidism and increased sepsis risk may be attributed to the altered immune responses characteristically seen in thyroid hormone deficiency. Thyroid hormones play a vital role in modulating immune function, and their deficiency can weaken immune defenses.¹⁵ Evidence suggests that reduced levels of thyroid hormones, such as free triiodothyronine (fT3) and free thyroxine (fT4), correlate with higher mortality rates in septic patients, indicating a potential link between hypothyroidism and increased postoperative infection susceptibility.¹⁶ This mechanism likely contributes to the elevated rates of sepsis and chronic device related infections, a 127% relative increase in our study, observed among patients with a hypothyroidism code in our cohort. While the absolute risk differences for these infectious complications were modest (0.8% for sepsis, 0.4% for device infection), the relative risk increases are noteworthy. Given the significant morbidity, mortality, and costs associated with sepsis and chronic spinal implant infections, even a small increase in risk warrants clinical attention. However, whether these specific risk increases justify widespread changes in preoperative screening or management protocols requires further investigation into the efficacy and cost-effectiveness of potential interventions.

In addition to perioperative complications, hypothyroidism’s effects on bone metabolism may offer a theoretical basis for exploring the potential link to long-term mechanical complications, such as PJK. Thyroid hormones are essential for maintaining bone density and strength; their deficiency can lead to reduced bone mineral density and structural integrity, rendering bones more susceptible to fractures and deformities.⁷ This disruption in bone metabolism may help explain the increased risk of proximal junctional kyphosis seen in patients with hypothyroidism. However, our finding of an increased risk of PJK (RR 1.42) associated with hypothyroidism should be interpreted with caution. As detailed in the Limitations section, diagnosing PJK accurately requires specific radiographic parameters not available in this administrative dataset Hence, the observed association which relies on the accuracy of ICD and CPT codes may benefit from future studies that also utilize definitive radiographic endpoints.

Interestingly, some studies have highlighted potential cardioprotective effects of hypothyroidism. Luther et al conducted a large-scale analysis of over 4 million patients in the National Inpatient Sample database (2004-2014), finding that hypothyroid patients undergoing both lumbar and cervical fusion exhibited increased rates of hematologic complications but demonstrated lower in-hospital mortality rates (lumbar fusion: OR 0.643; cervical fusion: OR 0.606) and reduced incidence of perioperative myocardial infarction (MI) in lumbar fusions (OR 0.851).¹¹ One possible explanation hypothesized by those authors was that the lower metabolic demand associated with hypothyroidism may lessen cardiovascular strain during surgery, thereby reducing the risk of perioperative myocardial infarction.¹⁵ However, these observational associations are complex to interpret and do not necessarily imply a protective effect; hypothyroidism itself carries known long-term cardiovascular risks. Furthermore, such database findings are susceptible to confounding. This limitation underscores the importance of further research and caution to clarify the complex relationship between hypothyroidism and perioperative cardiovascular events. Importantly, our propensity-matched analysis did not find significant differences in 45-day myocardial infarction or mortality rates between the hypothyroid and control groups.

Strengths and Limitations

Strengths of this study include a large sample size, propensity matching model, and cohort design that allowed us to examine time-to-event data. The TriNetX data utilized in this study encompasses over 100 million patients from over 100 healthcare organizations, providing a comprehensive dataset of demographic and clinical variables, thereby enhancing the study’s generalizability and applicability. Finally, to the best of our knowledge, this study is the first of its kind to evaluate long-term (2-year) mechanical outcomes in hypothyroid patients following lumbar fusion, expanding the scope of research beyond immediate postoperative outcomes.

This study has some notable limitations. As a retrospective observational analysis, this study evaluated association and not causation. Thus, prospective studies are needed to validate or refute the observed association between hypothyroidism and postsurgical complications.

Furthermore, a key limitation is the reliance on the administrative ICD-10-CM code E03.9 (“Hypothyroidism, unspecified”) to identify the exposure cohort. it inherently limits our understanding of the clinical specifics for individual patients. This broad code captures a heterogeneous group varying in disease etiology, duration, severity (eg, subclinical vs overt), and importantly, biochemical control status. Crucially, we lack data on Thyroid Stimulating Hormone (TSH), free T4, or free T3 levels, which are direct measures of thyroid function. This absence of granular hormonal data prevents us from exploring dose-response relationships (eg, does more severe hypothyroidism confer a greater risk?), distinguishing treated vs untreated patients accurately, or performing subgroup analyses based on biochemical control, potentially diluting the observed effect sizes, and limiting the precision of risk attribution.

Equally important is the limitation related to outcome ascertainment using administrative codes, particularly for PJK, PJF, and pseudoarthrosis. These mechanical complications are primarily identified through specific radiographic criteria, which may not be fully captured by ICD or CPT codes. As such, reliance on administrative codes alone may introduce some degree of misclassification—both false positives and false negatives—and potentially lead to an underestimation of true incidence. This should be considered when interpreting findings related to these outcomes.

Additionally, we were unable to discern the clinical severity or exact management of identified complications. The database also lacks information on patient compliance with thyroid medication, which may significantly impact surgical outcomes. Other limitations include the absence of patient-reported outcome measures (PROMs) that would better characterize functional recovery and quality of life, potential loss to follow-up for patients who sought care at non-participating institutions, limited granularity regarding specific surgical techniques beyond CPT codes (eg, number of levels fused, instrumentation details, bone graft types), information on bone mineral density, and lack of data regarding surgeon or institutional volume or experience. Variability in surgical techniques, approaches, and perioperative management protocols across different institutions could not be accounted for beyond the variables included in the matching. The TriNetX network aggregates data from many sources; specific details on data verification standards at the point of care within each contributing institution are not typically available to researchers.

Finally, due to the retrospective nature of this study, the potential for unmeasured confounding variables to affect the observed relationships must be considered. While we implemented propensity score matching to account for a range of observable demographic and comorbidity factors, retrospective studies using electronic health records are always susceptible to residual confounding from variables not routinely or consistently captured within EHR data. Examples of potentially unmeasured confounders in this context might include: the preoperative functional status of patients beyond broad comorbidity indices patient adherence to all medications including thyroid hormone replacement therapy, lifestyle factors (eg, detailed smoking history, nutrition) not reliably documented in EHRs, and subtle socioeconomic determinants of health. Furthermore, the strong association between hypothyroidism codes and metabolic factors (obesity, diabetes, hypertension) observed pre-matching meant that propensity score matching resulted in matched cohorts relatively enriched for these conditions. While enabling a valid comparison between groups, this may attenuate the observed effect size of hypothyroidism compared to a healthier baseline population and potentially limit the generalizability of the risk magnitude. Despite our rigorous matching, the potential for these and other unmeasured factors to contribute to the observed associations and potentially bias our findings cannot be fully eliminated.

Conclusion

This study demonstrates that patients with hypothyroidism are at an increased risk of both early and late postoperative complications following lumbar fusion. Sepsis, chronic device related infections, surgical wound complications, and proximal junctional kyphosis are all morbid entities that can significantly alter functional outcomes and impair a patient’s quality of life. Future research may consider refining preoperative management strategies, including determining the optimal timing for thyroid hormone replacement therapy, establishing minimum preoperative thyroid levels, and categorizing patients based on the severity of thyroid deficiency to better understand this intricate relationship.

Supplemental Material

Supplemental Material - Impact of Hypothyroidism on Short-Term and Long-Term Outcomes After Lumbar Fusion: A Nationwide Propensity-Matched Cohort Study

Supplemental Material for Impact of Hypothyroidism on Short-Term and Long-Term Outcomes After Lumbar Fusion: A Nationwide Propensity-Matched Cohort Study by Muhammad Waheed, MD, Abdul-Lateef Shafau, MD, Abdelrahman Diab, BS, Abdurrahman Ehsan, Sazid Hasan, MD, Omar Diab, BS, Bilal Butt, MD, and Ilyas Aleem, MD in Global Spine Journal

Footnotes

Author Contributions

Muhammad Waheed, MD: Conceptualization, study design, data collection, data analysis, manuscript writing, critical revision, final approval of the manuscript.

Abdul-Lateef Shafau, MD: Data collection, data analysis, literature review, manuscript drafting, critical revision, final approval of the manuscript.

Abdelrahman Diab, BS: Data collection, data analysis, literature review, manuscript drafting, critical revision, final approval of the manuscript.

Abdurrahman Ehsan: Data analysis, statistical interpretation, literature review, manuscript writing, critical revision.

Sazid Hasan, MD: Data collection, statistical analysis, manuscript drafting, critical revision.

Omar Diab, BS: Literature review, data organization, figure/table preparation, critical revision.

Bilal Butt, MD: Supervision, study design, critical revision, final approval of the manuscript.

Ilyas Aleem, MD: Supervision, conceptualization, study design, critical revision, final approval of the manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Ethical Statement

ORCID iDs

Muhammad Waheed

Abdul-Lateef Shafau

Ilyas Aleem

Data Availability Statement

The data supporting the findings of this study have been acquired from the TriNetX database. The datasets used and/or analyzed during this study are available from the corresponding author on reasonable request.*

Supplemental Material

Supplemental material for this article is available online.

References

Rajaee

Bae

Kanim

LEA

Delamarter

. Spinal fusion in the United States: analysis of trends from 1998 to 2008. Spine. 2012;37(1):67-76. doi:10.1097/BRS.0b013e31820cccfb

Reisener

Pumberger

Shue

Girardi

Hughes

. Trends in lumbar spinal fusion—a literature review. J Spine Surg. 2020;6(4):752-761. doi:10.21037/jss-20-502

Butrico

Meisel

Sage

. Patient comorbidities, their influence on lumbar spinal fusion surgery, and recommendations to reduce unfavorable outcomes. J Am Acad Orthop Surg. 2024;32(24):1115-1121. doi:10.5435/JAAOS-D-23-00312

Bassett

JHD

Williams

. Role of thyroid hormones in skeletal development and bone maintenance. Endocr Rev. 2016;37(2):135-187. doi:10.1210/er.2015-1106

Rashad

Samir

. Prevalence, risks, and comorbidity of thyroid dysfunction: a cross-sectional epidemiological study. Egypt J Intern Med. 2019;31(4):635-641. doi:10.4103/ejim.ejim_69_19

Vestergaard

Mosekilde

. Fractures in patients with hyperthyroidism and hypothyroidism: a nationwide follow-up study in 16,249 patients. Thyroid. 2002;12(5):411-419. doi:10.1089/105072502760143869

Dhanwal

. Thyroid disorders and bone mineral metabolism. Indian J Endocrinol Metab. 2011;15(Suppl 2):S107-S112. doi:10.4103/2230-8210.83349

Natori

Shimizu

Nagahama

Tanaka

. The influence of hypothyroidism on wound healing: an experimental study. Nihon Ika Daigaku Zasshi. 1999;66(3):176-180. doi:10.1272/jnms.66.176

Vakharia

Ameri

Niedzielak

Donnally

3rd Malloy

4th . Hypothyroidism increases 90-day postoperative complications in patients undergoing primary single level anterior cervical disectomy and fusion: a matched control analysis. J Spine Surg. 2018;4(2):274-280. doi:10.21037/jss.2018.05.26

10.

Buller

Rosas

Sabeh

Roche

McLawhorn

Barsoum

. Hypothyroidism increases 90-day complications and costs following primary total knee arthroplasty. J Arthroplast. 2018;33(4):1003-1007. doi:10.1016/j.arth.2017.10.036

11.

Luther

Perez-Roman

McCarthy

, et al. Incidence and clinical outcomes of hypothyroidism in patients undergoing spinal fusion. Cureus. 2021;13(8):e17099. doi:10.7759/cureus.17099

12.

Kim

Ahn

, et al. Risk factors for mechanical complications after fusion extension surgery for lumbar adjacent segment disease. Eur Spine J. 2024;33(1):324-331. doi:10.1007/s00586-023-07832-4

13.

Evaniew

Swamy

Jacobs

, et al. Lumbar fusion surgery for patients with back pain and degenerative disc disease: an observational study from the Canadian Spine Outcomes and Research Network. Glob Spine J. 2022;12(8):1676-1686. doi:10.1177/21925682221093615

14.

Vakharia

Ehiorobo

Mahmood

Roche

Mont

Razi

. Does hypothyroidism increase complications, lengths of stay, readmissions, and costs following primary 1- to 2-level lumbar fusion? Clin Spine Surg. 2020;33(10):E559-E562. doi:10.1097/BSD.0000000000001064

15.

Udovcic

Pena

Patham

Tabatabai

Kansara

. Hypothyroidism and the heart. Methodist DeBakey Cardiovasc J. 2017;13(2):55-59. doi:10.14797/mdcj-13-2-55

16.

Foks

Dudek

Polok

Nowak-Kózka

Fronczek

Szczeklik

. Thyroid hormones as potential prognostic factors in sepsis. Anaesthesiol Intensive Ther. 2019;51(3):205-209. doi:10.5114/ait.2019.88258

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB