Association of Bacterial Etiology With Rates of Reoperation in Patients With Septic Tenosynovitis of the Hand

The anatomy of the hand makes it uniquely sensitive to complications after bacterial infection. An intricate system of pulleys, tendon sheaths, compartments, and other closed systems can cause increases in pressure when inflamed, obstructing blood flow and predisposing patients to necrosis and potentially devastating complications.¹ Distension between the visceral and parietal layers of tendon sheaths can disrupt normal anatomic barriers and can also lead to cellulitis, erysipelas, fasciitis, and compartment syndrome.² Contractures, stiffness, and the need for more aggressive interventions, such as amputation, can result from improper or untimely treatment. While up to 80% of infections involve Staphylococcus aureus and Streptococcal species, there has been an increase in prevalence of methicillin-resistant S aureus (MRSA), with rates reported above 70%.^3-5 Patients with various comorbid conditions as well as those with injuries sustained from various mechanisms are prone to infection with specific pathogenic etiologies.

MRSA infections are more prevalent in patients who are immunocompromised, have diabetes, or are intravenous drug users,^6,7 whereas Eikenella species are more common in human-bite injuries,^8,9 and Pasteurella species are more common in infections caused by animal bites.^10-12 These conditions most frequently manifest as cellulitis and/or abscess formation that often require numerous surgical debridements.^13,14 Optimal management requires the identification of patients at risk for increased surgery to reduce complication rates. Indeed, the causative organism has been implicated as a predictor of complications after surgery and may be responsible for unplanned, repeated operations.¹⁵ Such research suggests that there may be a clinically significant difference in virulence and disease severity of these organisms.

Despite the frequent need for secondary surgery for hand infections, there is little published literature regarding pathogenic etiology that may predispose a patient to the need for a repeated operative procedure in cases of septic tenosynovitis. We hypothesize that bacterial etiology is associated with different rates of initial operation and reoperation in these patients. We also hypothesize that patients with polymicrobial infections will have higher incidence of reoperation than those who have infections with individual organisms.

Data and Patient Selection

A retrospective analysis was conducted on the 2001-2013 Nationwide Inpatient Sample (NIS) database. The NIS is a national database created by the Agency for Healthcare Research and Quality and maintained by the Healthcare Cost and Utilization Project (HCUP). The NIS approximates a 20% stratified sample of all discharges from US community hospitals, including specialty hospitals and academic medical centers.¹⁶ The NIS is the largest publicly available all-payer inpatient database in the US, and its utilization continues to increase due to its accessibility and validated methodology.^17-19

The unique design of the NIS requires specific methodological considerations that are detailed in the available online tutorials and documentation, which were reviewed before our analysis.²⁰ In particular, a change in sampling strategy took place beginning with 2012 data, requiring a different set of weights to be applied starting in 2012. Use of specific International Classification of Diseases, 9th Revision (ICD-9) codes is also necessary to gather accurate estimates. Following these recommendations, HCUP trend weights were applied for records before 2012, and discharge weights were applied for records from 2012 and beyond.

NIS was queried for cases of tenosynovitis using ICD-9 diagnostic codes 727.04 and 727.05. Pathogen cultures were also identified with ICD-9 diagnostic codes, and surgical intervention was determined by using ICD-9 procedural codes pertaining to the hand, including synovectomy, exploration, bursotomy, incision and drainage, fasciotomy/ectomy, myectomy, manipulation, irrigation, and suturing (Table 1). Due to the low numbers of isolated Candida, Sporotrichosis, and Cryptococcus infections (n = 26, 10, and 0, respectively), these etiologies were grouped into a “fungal” category for analysis. Reoperation was determined if the same ICD-9 procedural codes were listed multiple times in the same admission. Demographic variables of race, age, and sex were collected. Cases were excluded if they were missing demographic data. As this study did not involve human subjects, institutional review board approval was not needed as per institutional policy.

Statistical Analysis

χ² analysis was performed to determine variables that were significantly associated with operative intervention. Binary logistic regression was subsequently performed, accounting for demographic and significant variables, to determine predictors of initial operation and reoperation. HIV/AIDS and diabetes status were included in the initial analysis, as these are known to be predictors of adverse complications of hand infections.²¹ Outcomes of interest included initial surgery and need for additional surgery, which was defined as presence of a reoperation in the same patient. Descriptive data on demographics and pathogen etiology were recorded (Table 2). The relative proportion of each pathogen was calculated as 2 percentages: positive cultures if the organism was the only one cultured (the “isolated” percentage) and positive cultures regardless of whether it was the only one cultured or not. For example, 16.2% of all cultures grew streptococci only (ie, isolated), but streptococci were found in 21.9% of cultures overall.

In the multivariate logistic regression analysis, significance was defined as P < .003 after applying the Bonferroni correction, based on a total of 17 statistical comparisons.²² The Bonferroni Correction is a statistical method used when two or more analyses are being performed on the same sample of data.²² Significance was otherwise defined as P < .05.

A weighted total of 47,690 cases were included in the final analysis. Most patients were white men aged 30 to 50 years. The most common bacterial etiology was methicillin-sensitive Staphylococcu aureus (MSSA), which accounted for 50.8% of all positive cultures and was the single isolated pathogen in 46.3% of cases with positive cultures (Table 2). Only 7.5% of infections were reported as polymicrobial (Figure 1). There were no reported cases of Cryptococci infections. The highest rate of reoperation was found in patients with polymicrobial and fungal infections (30.5% and 33.3% of those who had initial surgery, respectively; Table 3).

Regression analysis showed that all bacterial and polymicrobial infections were significantly associated with rates of initial surgery for septic tenosynovitis that were higher than those in patients without positive cultures. Patients with diabetes also had higher likelihood of initial surgery. Patients who were non-white, female, and aged >61 years had lower likelihood of initial surgery (Table 4).

Rates of reoperation were higher in the following age groups: 30 to 50 years, 51 to 60 years, 61 to 79 years, and ≥80 years. Other factors associated with higher rates of reoperation included streptococci infection, unspecified Staphylococcus infection, MSSA, MRSA, fungal, and polymicrobial infections, and diabetes. Women had a lower likelihood of reoperation compared with men (Table 4). Mycobacterial and unspecified bacterial infections were not associated with a high rate of reoperation. HIV/AIDS status was not significantly associated with initial surgery or reoperation and was therefore not included in multivariable analysis.

Repeated debridements and other procedures are often needed to clear infections. Though many other reasons for repeat surgery have been described, the effect of infectious etiology on rate of secondary surgery is not completely understood. This study presents a large, nationally representative cohort of patients with septic tenosynovitis and compares rates of surgery based on pathogen etiology. The study found that etiology results in different rates of reoperation in these patients. Independent of race, age, and sex, patients with cultured Streptococcus or Staphylococcus species and polymicrobial infections are significantly more likely to have a reoperation compared with those without positive cultures. The greatest likelihoods of reoperation were seen in patients with fungal and polymicrobial infections, whereas MSSA was the overall most common pathogen identified. Though fungal infections were also associated with reoperation, they were not associated with initial operation.

Of those patients who had an initial operation, the rate of reoperation was 16.7%. By pathogen, the greatest rate of reoperation was among those who had a fungal (33.3%) or polymicrobial infection (30.5%). Due to the relatively low prevalence of isolated fungal infections in the sample (n = 36), the high rate of reoperation among polymicrobial infections should be considered more clinically relevant. By comparison, the lowest rate of reoperation of those who had initial surgery was among those with an unspecified staphylococcal infection at 12.9%. These data demonstrate the high rates of failure of initial treatment or predilection for multiple surgeries based on pathogen. These differences may be more relevant in patients with diabetes, as the data corroborate the existing literature that suggests diabetes is an important source of morbidity in hand infections.^6,23

The literature states that staphylococcal and polymicrobial infections are significant contributors to hand infections and morbidity after treatment, with positive Staphylococcus cultures found in 23% to 80% of hand infections.^3,4,9,24 Polymicrobial etiologies account for about 19% of hand infections.⁹ Though this study's data show different numbers (50.8% of cultures had MSSA, 20.1% had MRSA, and 7.5% had polymicrobial), these organisms clearly contribute a significant burden to the pathogenesis of hand infections. Coupled with high rates of reoperation, it may be necessary to plan more aggressive interventions in patients with these cultures.

The Infectious Diseases Society of America recommends the use of antibiotics for abscesses associated with extensive infections with concomitant drainage, if feasible.²⁵ Effective drainage of the hand, however, is often difficult due to its intricate anatomy and compartmental architecture. If allowed to progress, inflammation from septic tenosynovitis results in distension of the flexor sheath, which may lead to the spread of disease into the subcutaneous planes, further compromising blood flow, antibiotic penetration, and the intrinsic healing processes.²⁶ Consequently, more frequent and extensive debridements may be indicated in these patients.

This study postulates that the progression of disease differs by etiology and may therefore influence treatment. The microbiological literature suggests that Staphyloccus species tend to manifest as an infection with local sequelae, such as abscesses, whereas Streptococcus species are more likely to be associated with extensive conditions, such as necrotizing fasciitis or erysipelas.^25,27-29 Other studies, however, fail to report a clinical distinction between these infectious agents with regard to extent of infection, systemic signs, severity of infection, or number of surgical interventions.³⁰ As such, surgical treatment typically depends on a patient’s clinical symptoms rather than culture results because antibiotic coverage is often broad enough to cover the most common infectious organisms. Nonetheless, it is important to anticipate rates of surgery wherever possible to prevent dangerous sequelae of emergency procedures and the need for repeated, potentially avoidable, morbid surgery. The present study shows that bacterial etiology is associated with different rates of reoperation, and appropriate precautions should be taken in anticipation of surgical treatment of patients with positive cultures.

This study is limited by the fact that the extent of infection, causative trauma, or Michon classification stage of each case, which are known predictors of surgical intervention, is not known.^31,32 The antibiotic regimen or duration of antibiotics given to each patient is also not known. This may affect rates of surgery in this population, as those with delayed or inappropriate medical treatment may have been more likely to have a reoperation. It is also worth noting that not all possible pathogens are accounted for in the analysis as not all are present in the ICD-9 coding system. It is therefore possible that patients with infections other than those investigated may bias the results. Additionally, as the NIS database reports only inpatient discharge-level data, those patients treated in the outpatient setting are not represented.

In this study the most common isolated bacterial etiologies (Staphylococcus and Streptococcus species) and polymicrobial infections were frequently associated with rates of surgery, which corroborates reports in the existing literature.^33,34 Patients with septic tenosynovitis were more likely to have staphylococcal or streptococcal infections, and the highest rates of reoperation were seen in those with fungal or polymicrobial infections. Although the majority of patients were successfully treated initially with surgery, a reoperation rate of 16.7% suggests a need for close postsurgical follow-up.

Affiliations: Rutgers Health, New Jersey Medical School Department of Orthopaedics, Newark, NJ

Correspondence: Robert DalCortivo, BBA; bobdalcortivo@gmail.com

Disclosures: Michael M Vosbikian, MD, receives honorarium for content authorship from The Journal of Bone and Joint Surgery Clinical Classroom and is an editorial board member for ePlasty. The authors disclose no other relevant financial or nonfinancial interests.