Nationwide analysis of adult nongonococcal native septic knee arthritis hospitalizations from 2016 to 2022

Abstract

Background

Septic arthritis (SA) is a medical emergency associated with substantial morbidity and mortality. Although the knee is the most commonly affected joint in adults, contemporary national data describing characteristics, mortality, and healthcare utilization in knee-specific native SA are limited.

Methods

We conducted a retrospective study of adult hospitalizations for nongonococcal native knee SA in the United States using the National Inpatient Sample database from 2016 to 2022. Hospitalizations were identified using ICD-10 diagnostic codes, excluding gonococcal and prosthetic joint infections. Thus, the analytic cohort was restricted to native joint disease and did not include prosthetic joint infection. Weighted analyses were performed to generate national estimates. Multivariable logistic regression was used to identify factors independently associated with in-hospital mortality.

Results

Among approximately 205 million adult hospitalizations, 92,290 (0.045%) had native SA of the knee. In-hospital mortality was 2.5%. Patients who died were older, had longer lengths of stay, and incurred more than double the median hospital charges compared with survivors. In multivariable analysis, factors independently associated with higher odds of in-hospital death included respiratory failure (odds ratio [OR] 8.54), sepsis (OR 3.57), liver disease (OR 3.45), cerebrovascular disease (OR 2.27), acute kidney injury (OR 2.05), chronic kidney disease (OR 1.66), myocardial infarction (OR 1.60), heart failure (OR 1.55), and increasing age. Survivors demonstrated substantial postacute care needs, with 36% discharged to skilled nursing or intermediate care facilities.

Conclusions

In this nationally representative cohort, nongonococcal native septic knee arthritis showed substantial in-hospital mortality, morbidity, and healthcare utilization. Mortality was driven primarily by multiorgan dysfunction rather than demographic or socioeconomic factors. Early identification and aggressive management of high-risk patients may improve outcomes.

Septic arthritis (SA) is an infection of the synovial joint space that can rapidly cause cartilage destruction and death if not treated promptly. Importantly, in the absence of trauma or instrumentation, native joint SA often reflects hematogenous seeding during bacteremia rather than serving as the primary cause of sepsis, although an infected joint can remain a focus that complicates control of systemic infection.^1–4 Typically urgent joint drainage is required in combination with intravenous antibiotics for effective management.^1,2 A diagnosis of SA is established through synovial fluid aspiration demonstrating synovial leukocytosis, positive gram stain, and/or positive microbial culture.^1,2 From an etiological perspective, SA is predominantly bacterial in origin, with Staphylococcus aureus representing the most common causative pathogen in adults.^1,2,5–7 Conditions that predispose to SA include rheumatoid arthritis, diabetes mellitus, chronic kidney disease (CKD), chronic liver disease, and immunosuppression.^3,4 In adults, the knee is the most commonly affected joint.^2,5,6,8

SA constitutes a serious medical condition with notable consequences at the population level. Incidence rates in developed countries reach 4 to 10 cases per 100,000 people,¹ with an increased incidence in older individuals. Even with appropriate management, SA can be associated with substantial morbidity and mortality. Estimated mortality rates within the hospital setting range from 2% to 11%.^5,9–13 Early source control, involving joint drainage along with appropriate antibiotic therapy, remains essential to prevent joint destruction and to help control a persistent infectious focus in patients with concurrent bacteremia or sepsis.^1,2

Despite the substantial morbidity and mortality associated with native SA, contemporary large-scale epidemiological data remain limited. The most recent US national level analysis, conducted by Singh et al, reported approximately 13,000 any joint SA–related hospitalizations annually between 2009 and 2012.⁹ Although this earlier work provided valuable national estimates, it did not distinguish between affected joints; thus, knee-specific native SA, which accounts for roughly half of adult cases, has not been evaluated separately.

We conducted a retrospective study using the most recent version of the National Inpatient Sample (NIS) database to characterize the demographic, clinical, and economic burden of adult nongonococcal native septic knee arthritis hospitalizations in the US from 2016 to 2022. We also evaluated factors associated with in-hospital mortality and compared our findings with previously published cohorts to better understand disease burden, outcomes, and resource utilization.

METHODS

We performed a retrospective study of all adult native knee SA hospitalizations from 2016 to 2022 in the NIS database (http://www.hcup-us.ahrq.gov). The NIS database, part of the Healthcare Cost and Utilization Project (HCUP), is a 20% sample representative of all nonfederal acute-care hospitals in the US and is the largest all-payer hospitalization database in the country. Weighted analyses were performed to generate national estimates. The NIS sampling frame covers more than 97% of the US population. Diagnoses were identified by International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10), and surgeries were identified by ICD-10 Procedure Coding System (ICD10-PCS). To protect patient privacy, the HCUP data use agreement prohibits reporting hospitalization counts between 1 and 10; such values are indicated as NR (not reported) in all tables.

Inclusion and exclusion criteria and study variables

The study population consisted of all adult hospitalizations with a principal or secondary diagnosis of native knee SA, identified by ICD-10 codes (see Supplemental Table 1). Exclusion criteria included gonococcal SA and prosthetic joint infection. Therefore, the cohort should be interpreted as native joint knee SA and is not intended to represent prosthetic joint infection or all SA hospitalizations. Study variables included age, sex, race/ethnicity, length of stay (LOS), total hospital charges, in-hospital mortality, weekend admission status, median household income by zip code, discharge information, and medical insurance type. Median household income by zip code was divided into quartiles. In 2022, the US dollar values for each quartile were as follows: quartile 1 = $1–$55,999; quartile 2 = $56,000–$70,999; quartile 3 = $71,000–$93,999; and quartile 4 = >$94,000. Clinically relevant comorbidities and complications were identified using ICD-10 diagnosis codes listed in Supplemental Table 1, including acute kidney injury (AKI), central line infection, delirium, electrolyte abnormalities, endocarditis, necrotizing fasciitis, opiate use, osteomyelitis, pneumonia, respiratory failure, sepsis, and venous thromboembolism. To provide a comprehensive assessment of patient complexity, we also examined individual components of the Charlson Comorbidity Index, including AIDS, cancer, cerebrovascular accident/transient ischemic attack (CVA/TIA), CKD, chronic obstructive pulmonary disease (COPD), diabetes mellitus, dementia, hemiplegia, heart failure, inflammatory arthritis or connective tissue disease (CTD), liver disease, myocardial infarction (MI), peptic ulcer disease, and peripheral vascular disease.

Statistical analysis

Analysis was performed using STATA, version 18 (StataCorp, College Station, Texas, USA). Baseline characteristics were summarized using descriptive statistics. Univariable logistic regression analyses were used to calculate unadjusted odds ratios (ORs) for in-hospital death. All regression analyses were conducted on the weighted sample and were limited to hospitalizations with a recorded mortality status. Variables with P values ≤ 0.2 in univariable analysis were included in the multivariable logistic regression model. P values ≤ 0.05 were considered statistically significant in the multivariable analysis.

Outcomes

The study assessed (1) factors associated with in-hospital mortality among patients hospitalized with nongonococcal native SA of the knee; (2) demographic characteristics, hospitalization metrics including LOS, and total hospital charges; and (3) discharge disposition and prevalence of clinically relevant complications and comorbidities.

RESULTS

We identified 205 million adult hospitalizations in the 2016–2022 NIS database. Among these, 92,290 hospitalizations (0.045%) involved a principal or secondary diagnosis of native joint nongonococcal knee SA. Of these, 89,875 hospitalizations (97.4%) were discharged alive, while 2320 (2.5%) ended in in-hospital death (Table 1). Mortality status was missing for 95 hospitalizations.

Table 1.

Characteristics of adult nongonococcal native knee septic arthritis from the National Inpatient Sample, 2016 to 2022

Characteristics	Total cohort (n = 92,290*)	Discharged alive (n = 89,875)	In-hospital death (n = 2320)	P value
Age (years): Median (IQR)	60 (47–71)	59 (47–71)	68 (59–78)	<0.001
Male	65.5%	65.4%	67.0%	0.48
White	66.7%	66.8%	61.2%	0.01
African American	14.0%	14.0%	16.6%	0.11
Hispanic	10.2%	10.2%	11.2%	0.47
Asian/Pacific Islander	2.0%	2.0%	4.5%	<0.001
Native American	1.5%	1.5%	1.1%	0.47
Other race	2.3%	2.3%	2.4%	0.88
Income quartile 1	30.8%	30.8%	28.7%	0.32
Income quartile 2	26.1%	26.0%	27.8%	0.38
Income quartile 3	22.5%	22.4%	24.4%	0.33
Income quartile 4	18.1%	18.1%	17.0%	0.55
Medicare	44.6%	44.2%	59.7%	<0.001
Medicaid	20.5%	20.6%	17.0%	0.06
Private insurance	25.1%	25.3%	16.4%	<0.001
Self-pay	5.4%	5.4%	2.8%	0.01
No charge	0.5%	0.5%	NR	0.81
Other insurance	3.8%	3.8%	3.7%	0.91
Weekend hospital admission	22.0%	21.9%	25.0%	0.11
Clinically relevant complications and comorbidities in native septic arthritis
Acute kidney injury	23.9%	22.7%	67.9%	<0.001
Cancer	7.3%	7.2%	10.6%	0.01
Central line infection	0.9%	0.9%	2.0%	0.02
Cerebrovascular accident/TIA	1.1%	1.0%	8.4%	<0.001
Chronic kidney disease	20.9%	20.3%	43.8%	<0.001
COPD	10.2%	10.1%	12.9%	0.05
Delirium	1.5%	1.4%	6.0%	<0.001
Dementia	3.7%	3.6%	7.3%	<0.001
Diabetes mellitus	35.9%	35.7%	42.5%	0.01
Electrolyte abnormalities	35.9%	35.2%	60.3%	<0.001
Endocarditis	4.0%	3.8%	11.9%	<0.001
Heart failure	4.5%	4.3%	14.2%	<0.001
Hemiplegia	0.3%	0.3%	1.9%	<0.001
HIV	0.7%	0.7%	NR	0.18
Inflammatory arthritis/CTD	17.6%	17.8%	12.3%	0.002
Liver disease	5.2%	4.8%	20.9%	<0.001
Myocardial infarction	2.1%	1.8%	12.3%	<0.001
Necrotizing fasciitis	0.9%	0.8%	2.6%	<0.001
Opiates	11.3%	11.5%	5.0%	<0.001
Osteomyelitis	11.4%	11.5%	8.6%	0.06
Peptic ulcer disease	2.7%	2.7%	5.0%	0.003
Peripheral vascular disease	4.1%	4.1%	5.0%	0.35
Pneumonia	1.7%	1.5%	10.6%	<0.001
Respiratory failure	9.6%	8.1%	65.3%	<0.001
Sepsis	29.2%	28.0%	78.9%	<0.001
Venous thromboembolism	4.0%	3.8%	9.1%	<0.001

COPD indicates chronic obstructive pulmonary disease; CTD, connective tissue disease; HIV, human immunodeficiency virus; IQR, interquartile range; NR, not reported; TIA, transient ischemic attack.

^*Ninety-five cases had missing mortality data.

Patients who died were older (median age 68 vs 59 years, P < 0.001), had longer LOS (median 10 vs 7 days, P < 0.001), incurred significantly higher total hospital charges ($154,593 vs $63,544, P < 0.001), were more likely to have Medicare (59.7% vs 44.2%, P < 0.001), and were less likely to have private insurance (16.4% vs 25.3%, P < 0.001). No significant differences were observed in income quartiles. Posthospital disposition of survivors showed that 36% were discharged to skilled nursing/intermediate care facilities (SNF/ICF), 27.8% received home health care services, and 3.8% left against medical advice (Table 2).

Table 2.

Hospital disposition and cost characteristics in native septic knee arthritis, National Inpatient Sample, 2016 to 2022

Disposition and cost characteristics	Discharged alive (n = 89,875)	In-hospital death (n = 2320)	P value
Discharged to SNF or ICF	36.0%	–	–
Discharged with home health care	27.8%	–	–
Left against medical advice	3.8%	–	–
LOS (days): Median (IQR)	7 (4–12)	10 (4–19)	<0.001
Total charges: Median (IQR)	$63,544 (36,868–116,561)	$154,593 (69,249–322,079)	<0.001

ICF indicates intermediate care facility; IQR, interquartile range; LOS, length of stay; SNF, skilled nursing facility.

Several clinically relevant complications and comorbidities from the Charlson index were significantly more frequent among patients who died during the hospitalization (Table 3), including AKI (67.9% vs 22.7%, P < 0.001), cancer (10.6% vs 7.2%, P = 0.005), central line infection (2.0% vs 0.9%, P = 0.014), CVA/TIA (8.4% vs 1.0%, P < 0.001), CKD (43.8% vs 20.3%, P < 0.001), COPD (12.9% vs 10.1%, P = 0.045), delirium (6.0% vs 1.4%, P < 0.001), dementia (7.3% vs 3.6%, P < 0.001), diabetes mellitus (42.5% vs 35.7%, P = 0.003), electrolyte abnormalities (60.3% vs 35.2%, P < 0.001), endocarditis (11.9% vs 3.8%, P < 0.001), heart failure (14.2% vs 4.3%, P < 0.001), hemiplegia (1.9% vs 0.3%, P < 0.001), liver disease (20.9% vs 4.8%, P < 0.001), MI (12.3% vs 1.8%, P < 0.001), necrotizing fasciitis (2.6% vs 0.8%, P < 0.001), peptic ulcer disease (5.0% vs 2.7%, P = 0.003), pneumonia (10.6% vs 1.5%, P < 0.001), respiratory failure (65.3% vs 8.1%, P < 0.001), sepsis (78.9% vs 28.0%, P < 0.001), and venous thromboembolism (9.1% vs 3.8%, P < 0.001). The following were lower in the deceased patients: inflammatory arthritis/connective tissue disease (12.3% vs 17.8%, P = 0.002) and opiate use disorder (5.0% vs 11.5%, P < 0.001),

Table 3.

Univariable analysis for in-hospital death among native septic knee arthritis hospitalizations, National Inpatient Sample, 2016 to 2022

Variable	OR	P value	95% CI
Age	1.04	<0.001	1.030–1.043
Male	1.07	0.48	0.881–1.311
White	0.78	0.01	0.649–0.945
African American	1.23	0.11	0.957–1.571
Hispanic	1.11	0.47	0.829–1.497
Asian/Pacific Islander	2.36	<0.001	1.503–3.704
Native American	0.73	0.486	0.304–1.743
Other race	1.05	0.88	0.572–1.921
Income quartile 1	0.90	0.32	0.735–1.105
Income quartile 2	1.09	0.38	0.895–1.340
Income quartile 3	1.11	0.33	0.899–1.375
Income quartile 4	0.93	0.55	0.728–1.185
Medicare	1.87	<0.001	1.548–2.252
Medicaid	0.79	0.06	0.619–1.012
Private insurance	0.58	<0.001	0.450–0.741
Self-pay	0.50	0.02	0.288–0.873
No charge	0.84	0.81	0.213–3.332
Other insurance	0.97	0.91	0.597–1.582
Admission on weekend	1.19	0.11	0.960–1.464
Acute kidney injury	7.18	<0.001	5.900–8.747
Cancer	1.53	0.006	1.132–2.057
Central line infection	2.29	0.02	1.163–4.504
Chronic kidney disease	3.05	<0.001	2.535–3.673
COPD	1.32	0.046	1.005–1.734
CVA/TIA	9.50	<0.001	6.629–13.609
Delirium	4.63	<0.001	3.098–6.916
Dementia	2.12	<0.001	1.484–3.032
Diabetes mellitus	1.33	0.003	1.104–1.600
Electrolyte abnormality	2.80	<0.001	2.320–3.373
Endocarditis	3.45	<0.001	2.568–4.625
Heart failure	3.70	<0.001	2.808–4.864
Hemiplegia	6.44	<0.001	3.155–13.164
HIV	0.28	0.21	0.039–2.019
Inflammatory arthritis/CTD	0.65	0.002	0.490–0.854
Liver disease	5.30	<0.001	4.190–6.701
Myocardial infarction	7.53	<0.001	5.592–10.152
Necrotizing fasciitis	3.22	<0.001	1.784–5.81
Opiate	0.40	<0.001	0.265–0.614
Osteomyelitis	0.73	0.06	0.525–1.01
Peptic ulcer disease	1.89	0.003	1.234–2.892
Peripheral vascular disease	1.22	0.35	0.8–1.871
Pneumonia	7.89	<0.001	5.732–10.862
Respiratory failure	21.27	<0.001	17.449–25.934
Sepsis	9.62	<0.001	7.663–12.088
Venous thromboembolism	2.49	<0.001	1.795–3.443

CI indicates confidence interval; COPD, chronic obstructive pulmonary disease; CTD, connective tissue disease; CVA, cerebrovascular accident; OR, odds ratio; TIA, transient ischemic attack.

In multivariable logistic regression (Table 4), the following variables were independently associated with greater odds of in-hospital death (listed from highest to lowest ORs): respiratory failure (OR 8.54; 95% CI 6.665–10.930), sepsis (OR 3.57; 95% CI 2.735–4.661), liver disease (OR 3.45; 95% CI 2.593–4.596), CVA/TIA (OR 2.27; 95% CI 1.388–3.702), AKI (OR 2.05; 95% CI 1.611–2.599), CKD (OR 1.66; 95% CI 1.318–2.097), MI (OR 1.60; 95% CI 1.058–2.409), heart failure (OR 1.55; 95% CI 1.113–2.164), and age (OR 1.04; 95% CI 1.026–1.050). Inflammatory arthritis/CTD (OR 0.64; 95% CI 0.469–0.865), COPD (OR 0.68; 95% CI 0.493–0.944), and opiate use disorder (OR 0.60; 95% CI 0.371–0.983) were associated with lower odds of in-hospital death.

Table 4.

Multivariable analysis for in-hospital death among native septic knee arthritis hospitalizations in the National Inpatient Sample, 2016 to 2022

Variable	OR	P value	95% CI
Age	1.04	<0.001	1.026–1.050
White	0.92	0.58	0.688–1.233
African American	1.26	0.22	0.875–1.811
Asian/Pacific Islander	1.63	0.10	0.905–2.931
Medicare	0.76	0.37	0.418–1.383
Medicaid	1.10	0.77	0.596–2.014
Private insurance	0.83	0.55	0.449–1.528
Self-pay	0.86	0.73	0.370–2.013
Admission on weekend	1.02	0.88	0.796–1.304
Acute kidney injury	2.05	<0.001	1.611–2.599
Cancer	1.26	0.19	0.891–1.777
Central line infection	0.98	0.97	0.438–2.214
Chronic kidney disease	1.66	<0.001	1.318–2.097
COPD	0.68	0.02	0.493–0.944
CVA/TIA	2.27	0.001	1.388–3.702
Delirium	1.57	0.09	0.932–2.652
Dementia	1.12	0.64	0.708–1.756
Diabetes mellitus	0.96	0.69	0.767–1.190
Electrolyte abnormality	1.06	0.60	0.847–1.334
Endocarditis	1.41	0.11	0.925–2.155
Heart failure	1.55	0.01	1.113–2.164
Hemiplegia	1.27	0.54	0.588–2.745
Inflammatory arthritis/CTD	0.64	0.004	0.469–0.865
Liver disease	3.45	<0.001	2.593–4.596
Myocardial infarction	1.60	0.03	1.058–2.409
Necrotizing fasciitis	1.11	0.79	0.522–2.363
Opiate	0.60	0.04	0.371–0.983
Osteomyelitis	0.73	0.09	0.512–1.054
Peptic ulcer disease	0.88	0.63	0.536–1.458
Pneumonia	1.22	0.32	0.823–1.811
Respiratory failure	8.54	<0.001	6.665–10.930
Sepsis	3.57	<0.001	2.735–4.661
Venous thromboembolism	1.04	0.87	0.676–1.585

CI confidence interval; COPD, chronic obstructive pulmonary disease; CTD, connective tissue disease; CVA, cerebrovascular accident; OR, odds ratio; TIA, transient ischemic attack.

DISCUSSION

This study represents the largest analysis of native knee SA to date, encompassing more than 90,000 hospitalizations captured in a nationally representative US database from 2016 to 2022. The observed in-hospital mortality rate was 2.5%, which is notably lower than rates reported in several prior studies.^5,10–13 The lower mortality in this analysis compared to historical cohorts may reflect restriction of the cohort to native knee SA cases only, advances in hospital care, and/or inability to account for posthospital mortality. Prior longitudinal studies have reported long-term mortality rates of 17.5% to 28.3%.⁸

Respiratory failure and sepsis

Respiratory failure and sepsis showed the strongest associations with mortality in our cohort, underscoring the critical impact of both severe pulmonary compromise and systemic infection on SA patient outcomes. More than two-thirds of patients who died had respiratory failure, compared with only 8.1% of survivors, and this condition was associated with 8.54-fold higher odds of in-hospital death. Nearly 80% of those with fatal outcomes were admitted with sepsis or developed sepsis during their hospital stay. Importantly, the directionality between SA and sepsis cannot be determined using the NIS. In adults without preceding trauma or instrumentation, native joint SA commonly arises from hematogenous seeding during bacteremia; however, an infected joint can also function as a persistent focus that complicates treatment of systemic infection or increases recurrence risk if source control is delayed. Thus, our sepsis findings should be interpreted as an association with severe systemic infection, not as evidence that knee SA necessarily caused sepsis. These results are consistent with existing studies on the risk of sepsis in SA and further establish the devastating consequences of systemic infection.⁸

Renal and hepatic disease

Nearly two-thirds of patients who died had AKI during their admission, compared to less than one-quarter of survivors. AKI was strongly associated with in-hospital mortality (OR 2.05). CKD was also associated with higher odds of death (OR 1.66), occurring in 43.8% of nonsurvivors versus 20.3% of survivors. These associations should be interpreted cautiously and should not be taken to imply a direct renal-mediated mechanism of mortality. AKI may reflect the overall severity of acute illness, sepsis-associated end-organ dysfunction, or multiorgan failure rather than a causal effect of kidney injury itself. Similarly, although CKD and end-stage renal disease are recognized markers of baseline comorbidity and infection susceptibility, our administrative dataset could not determine whether uremia, immune dysfunction, renal replacement therapy, or severity of acute illness drove the observed association with mortality. Consistent with our findings, prior studies have demonstrated that impaired renal function is associated with increased mortality in SA.⁸

Liver disease was strongly associated with mortality in our analysis (OR 3.45), occurring in 20.9% of nonsurvivors compared with 4.8% of survivors. These findings echo those of Schindler et al, who identified chronic liver disease as an important predictor of poor outcomes in SA.⁸ The excess risk is plausibly mediated by immune dysfunction and coagulopathy associated with hepatic impairment.^3,4

Cardiac and cerebrovascular disease

Cardiac and cerebrovascular conditions/complications were associated with elevated mortality, and these findings are supported by several existing studies on SA.^8,10,12 Within this broader category of cardiovascular conditions, specific diagnoses demonstrated particularly strong associations. Concurrent heart failure had 1.55-fold higher odds of death, with a significantly greater incidence among nonsurvivors (14.2%) compared to survivors (4.3%). Additionally, MI had a 1.60-fold increased odds of death, and its prevalence was nearly sevenfold higher among deceased patients. CVA/TIA was also associated with in-hospital death (OR 2.27). Its prevalence was markedly higher among nonsurvivors (8.4%) compared with survivors (1.0%).

Conditions with protective associations: inflammatory arthritis/CTD, COPD, and opiates

Underlying inflammatory arthritis/CTD was present in 17.6% of knee SA hospitalizations but was associated with lower odds of in-hospital death (OR 0.64). Our findings align closely with a large Korean national study, which demonstrated rheumatologic disease to be associated with reduced in-hospital and short-term mortality in septic knee arthritis (hazard ratio 0.61 for in-hospital mortality and 0.76 for 30- and 90-day mortality).¹⁰ A study in Germany similarly found that rheumatologic diseases did not confer an increased mortality risk; the odds ratio for death was 3.14 (95% CI 0.18–1.81), indicating no significant association.⁸ Thus, while inflammatory arthritis/CTD is an established risk factor for SA, available data do not show that it independently drives in-hospital mortality.

Our study revealed that COPD was associated with lower odds of in-hospital death (OR 0.68), with a similar prevalence among deceased patients (12.9%) and survivors (10.1%). The basis for this protective association remains unclear.

Opiate use was associated with lower in-hospital mortality (OR 0.60), and documented use was less common among nonsurvivors (5.0%) compared with survivors (11.5%). Intravenous drug use is a well-recognized risk factor for SA, but ICD-10 coding does not differentiate the route of opiate use, and substance use disorders are often underreported in the NIS database. Therefore, this seemingly protective association should be interpreted with caution.

Racial/ethnicity, gender, income and insurance

The majority of native knee SA hospitalizations involved White individuals (66.7%), followed by African American (14.0%), Hispanic (10.2%), Asian/Pacific Islander (2.0%), and Native American (1.5%) groups. Socioeconomic distribution was skewed toward lower income levels, with nearly one-third of hospitalizations in the lowest income quartile (30.8%) and progressively fewer in higher quartiles (26.1%, 22.5%, and 18.1%). Male hospitalizations comprised approximately two-thirds of the cohort, reflecting a notable gender predominance. Notably, multivariable analysis did not demonstrate race, income quartile, insurance type, or gender to be independent predictors of in-hospital mortality.

Economic and healthcare burden

Our analysis revealed the large economic cost of SA. Even among survivors, resource utilization was substantial, with a median hospitalization duration of 7 days and median hospital charge of $63,544. More than one-third of survivors were discharged to SNF/ICF, and an additional quarter were discharged with home health care, highlighting the significant postacute care needs in this population.

Strengths and weaknesses

The main strength of this study is the large sample size. With more than 90,000 hospitalizations, it is the largest native knee SA series reported to date. The native knee–specific focus reduces the heterogeneity seen in other studies. Furthermore, by evaluating each Charlson Comorbidity Index component separately, this study provides a more detailed clinical profile than prior national analyses.

There are several important shortcomings to this study, including its retrospective design and reliance on administrative coding. Because prosthetic joint infections were excluded, this study does not capture cases in which indwelling hardware and biofilm-associated infection may contribute to greater morbidity, mortality, surgical complexity, and resource utilization; therefore, the findings are applicable to native knee SA hospitalizations and should not be generalized to prosthetic joint infection or all SA cases. Additionally, the NIS database does not reliably capture microbiologic data, cause-specific mortality, timing of antibiotic administration, antecedent knee procedures or trauma, or clinical decision-making rationale. Therefore, we could not determine whether SA was the primary driver of death, a complication of bacteremia or sepsis, or a coexisting condition during severe systemic illness. Similarly, although prosthetic joint infections were excluded, we could not assess whether preceding arthrocentesis, intra-articular injection, arthroscopy, or trauma contributed to joint-space seeding. Because laboratory values, uremia severity, dialysis or renal replacement therapy timing, organ-failure severity scores, and the temporal sequence of AKI, sepsis, and death are not available in the NIS, this analysis could not determine whether AKI or CKD contributed causally to mortality or served as markers of critical illness and baseline comorbidity. Given the retrospective design of this study, surgical drainage procedures were not included in the multivariable analysis due to potential selection bias; clinically unstable patients may not have been considered suitable candidates for surgery. The absence of a dedicated ICD-10 code for intravenous drug use necessitated the use of general opiate use as a surrogate marker, which may not fully capture the intended population and resulted in the protective odds ratio. Importantly, the NIS is a hospitalization-level database and does not contain unique patient identifiers; therefore, these data reflect hospitalizations for SA rather than unique individuals. As a result, recurrent admissions cannot be distinguished. Finally, the database lacks longitudinal follow-up, preventing assessment of important postdischarge outcomes such as functional recovery or long-term mortality.

CONCLUSION

In this nationally representative analysis of adult hospitalizations with nongonococcal native knee SA in the United States from 2016 to 2022, we found that although the condition was rare (0.045% of all adult hospitalizations), it carried a substantial burden of morbidity, mortality, and healthcare utilization. In-hospital mortality occurred in 2.5% of cases, with nonsurvivors experiencing significantly longer hospital stays and more than double the hospitalization charges. Additionally, survivors had high rates of discharge to SNF/ICF or need for home health services. Mortality was driven primarily by acute multiorgan dysfunction and comorbidities of the Charlson index rather than demographic or socioeconomic factors. Factors independently associated with higher in-hospital mortality were respiratory failure, sepsis, liver disease, CVA/TIA, AKI, CKD, MI, heart failure, and age.

Beyond timely antibiotics and joint drainage, early recognition of high-risk clinical profiles, particularly those with respiratory, hepatic, renal, or cardiovascular compromise, represents an opportunity to improve outcomes in patients hospitalized with native knee SA.

Disclosure statement/Funding

The authors report no funding or conflicts of interest.