Spread of Hand Infection According to the Site of Entry and Its Impact on Treatment Decisions

Patients and Methods

The data of all patients with hand infections who were admitted consecutively from 2007 to 2015 to our Department of Hand Surgery were reviewed retrospectively. All data were obtained from the medical databases of the electronic patient information system SAP (Walldorf, Germany). The inclusion criteria are summarized in Table 1.

Patient demographics and clinical data were analyzed for age, gender, and dates of injury, admission, and discharge. We extracted cases of hand infection by manually researching each case record. We evaluated age, existing illnesses, the microbiologic profile, laboratory results (C-reactive protein [CRP], white blood cell [WBC] count), the location of the wound and spread of infection as derived from patient charts, surgical records, and photographic documentation. Anatomic areas (finger, thumb, thenar, midpalmar, hypothenar, dorsal hand) were categorized as the portal of entry. The same process applied for the categorization of the spread of infection according to the literature. Nine spaces were included as hand compartments to which infections could spread: Thenar space, Parona space, midpalmar space, dorsal space, interdigital web space, Loge-de-Guyon, carpal tunnel, radiocarpal joint, and tendon sheaths.

The results are expressed as mean ± standard deviation. N numbers record the results per year if not otherwise indicated. For comparison of multiple experimental groups, one-way analysis of variance (ANOVA) was performed where indicated. The Dunnett multiple comparison post-test was performed following multiple comparison with ANOVA when the overall p value was <0.05.

Results

Co-morbidities and immunosuppression

Co-morbidities and advanced age correlated with the severity of the spread of infection (Fig. 1). Compared with young and healthy patients, significantly more compartments were affected in patients suffering from diabetes mellitus or some sort of immunosuppression or vascular disease (diabetes mellitus p = 0.027; immunosuppression p = 0.0259; and PAD p = 0.0345) (Fig. 1). Arterial hypertension did not affect the severity of hand infection significantly (p = 0.6098) (Fig. 1A).

OPEN IN VIEWER

FIG. 1.

Dunnett multiple comparison post-test was performed after performing multiple comparisons with analysis of variance when overall p value was <0.05. Error bars represent standard deviation. n indicates results per year. Each group was compared with young (1A) and healthy (1B) patients. ***p < 0.001; **p < 0.01; *p < 0.05.

Mechanisms of injury

We observed 15 mechanisms of injury: Human, dog, and cat bites, burns, penetrating injury, acute paronychia, intravenous drug abuse, minor trauma, operations, venous puncture, insect bite, foreign body, contused lacerated wounds, non-traumatic, and others. The number of compartments affected was highest in animal and human bite injuries (Fig. 2). Hand infection caused by cat bites affected an average of 3.3 compartments, dog bites 3.0, and human bites 2.8.

OPEN IN VIEWER

FIG. 2.

Cause of infections according to number of compartments involved.

Laboratory markers and microbiology findings

As an indicator of infection, CRP concentrations showed that the more compartments of the hand that were infected, the higher the values were (Fig. 3A). The CRP value increased from 16.1 mg/dL with ≤1 infected compartment to more than 100 mg/dL for more than four additional affected compartments. The mean CRP value was elevated to 41 mg/dL while the mean WBC count was 9.7 × 10³/mcL. Comparison of groups showed the following p values: CRP <20 mg/dL versus 20–39; p = 0.022; CRP 20–39 mg/dL versus 40–59; p = 0.012; CRP 40–59 mg/dL versus 60–79; p = 0.004; CRP 60–79 g/dL versus 80–99; p = ns; and CRP 80–99 mg/dL versus >100; p = <0.0001. There was no significant association between the WBC count and the severity of hand infection (Fig. 3B).

OPEN IN VIEWER

FIG. 3.

Dunnett multiple comparison post-test was performed after performing multiple comparisons with analysis of variance when overall p value was <0.05. ***p < 0.001; **p < 0.01; *p < 0.05. Error bars represent standard deviation. n indicates results per year. C-reactive protein values (3A) and white blood cell counts (3B) were analyzed according to number of compartments involved.

Mixed infections with as many as four bacterial species affected significantly more compartments per identified species (one versus two p = 0.0003; two versus three p = 0.013; three versus four p = 0.034). Mixed infections with more than four species did not increase the number of infected compartments further (four versus five p = 0.725; five versus six p > 0.999) (Fig. 4).

OPEN IN VIEWER

FIG. 4.

Dunnett multiple comparison post-test was performed after performing multiple comparisons with analysis of variance when overall p value was <0.05. Error bars represent standard deviation. n indicates results per year. Patient groups were compared according to number of bacterial. species isolated. ***p < 0.001; **p < 0.01; *p < 0.05.

Effect on hand compartments

The primary focus of the current study was the pathway by which the hand infection spreads from its origin. First, the number of compartments affected in the hand was analyzed by entry location (Fig. 5). Especially, wounds of the fingers had a tendency to cross anatomic borders. Infection of the index and small fingers affected the most compartments (Fig. 5). Regardless of the wound location, the tendon sheaths were affected in 50% of all patients and were the most common infection pathway. Finger infections spread in 20% each to the palm or dorsum of the hand, and 30% showed involvement of the interdigital web spaces. Thumb infections involved the tendon sheaths especially in 43% and the thenar space in 20% (Fig. 6). The infections of the hypothenar region also often showed involvement of interdigital web spaces (45%). The transmission pattern of infection after injuries of the midhand showed affliction of the tendon sheath in 49% and the dorsal spaces in 43% (Fig. 6B). Thenar infections spread especially into tendon sheaths in 25% (Fig. 6B).

OPEN IN VIEWER

FIG. 5.

Number of affected compartments according to entry site.

OPEN IN VIEWER

FIG. 6.

Pattern of spread according to entry site.

Discussion

The principles for the surgical management of hand infections are well described in the literature [6]. Successful management is mostly a consequence of the surgeon's experience, the severity of the infection, the microbiologic features, and individual risk factors [7-9]. Hand infections often originate from, or expand into, multiple bradytrophic compartments. Detailed knowledge of the severity and possible spread of the infection is essential for adequate treatment.

To our knowledge, no systematic investigation of spread, microbiologic patterns, risk factors, and laboratory values of hand infection according to its severity has been published. Our study confirms that, compared with healthy patients, patients with co-morbidities as diabetes mellitus, chronic vascular disease, or immunosuppression as well as older age have a higher risk of suffering severe hand infections (see Fig. 1). These findings are in line with published data [10]. In particular, indications for surgical treatment of hand infections in patients with co-morbidities should be more permissive. Complications and severe progression of infections are caused not only by particularly virulent pathogens and co-morbidities: They are the result of different internal and external factors. The timing of risk factor detection and its influence on treatment start time also is an important issue [9]. Besides the injuries themselves, being neglected by the patient and missed diagnosis by physicians can lead to significant delay in treatment. This delay can result in severe complications such as joint stiffness, necrosis, or even amputation [5,9]. In our previous study, we revealed that late presentation of patients with hand infections to hand surgeons can lead to a longer inpatient stay and a higher number of necessary operations [9]. Therefore, primary treating physicians should always take hand infections seriously and present the patient to a hand surgeon at an early stage [9].

In the current study, dog, cat, and human bite injuries were the most common causes of severe hand infections (see Fig. 2). Because of the microbiologic spectrum and wound depth, human as well as animal bites can lead to severe infections. Rothe and colleagues described that 10%–20% of bite wounds become infected, including 30%–50% of cat bites, 5%–25% of dog bites, and 20%–25% of human bites [11]. The surgical treatment of bite injuries is based on excision of the bite channel and removal of contaminated and necrotic tissue, mechanical reduction of the burden of pathogens, exclusion of injuries of functional structures, infection prophylaxis using broad-spectrum antibiotics, immobilization of fingers or the entire hand and, if indicated, administration of tetanus vaccine [2]. Complications such as osteomyelitis, septic arthritis, deep space infections, and stiffness are common. Especially in human bite infections, a delay in initial treatment, inadequate debridement, and inappropriate wound closure timing are confounding factors [12].

In addition to the clinical signs, radiologic imaging plays a major role in the diagnosis of hand infections [13]. Ultrasound scans offer a good assessment of soft tissues, but they require expertise that may not be available at a particular time. Besides, ultrasound cannot assess the underlying skeletal structures [13]. Some forms of hand infections require initial evaluation with radiographs to detect foreign bodies, osteitis, or fractures [3,13]. However, X-ray examinations provide limited assessment of the regional soft tissues [13]. Magnetic resonance imaging (MRI) enables better characterization of hand and wrist infections, particularly their soft tissue extent, but it usually is not available overnight and is time consuming [13].

Computed tomography (CT) has three-dimensional (3D) capabilities and is a useful tool in the emergency setting for assessment of advanced superficial and deep hand infections because of its rapid image acquisition and wide availability [13]. But even a CT scan offers only a late diagnosis of hand infections. An early diagnosis and initial assessment of the severity of the infection remain dependent on clinical signs, laboratory values, and, especially, the clinical experience of the physician. Some differential diagnoses are gout, pseudo-gout and acute hydroxyapatite deposition, pyogenic granuloma, acute non-infectious tenosynovitis, spider bites, inflammatory arthritis such as rheumatoid arthritis, foreign-body reactions, and paint gun injuries. These hand disorders also cause inflammation and edema in soft tissue and thus mimic hand infections. So they are difficult to distinguish from hand infections by diagnostic imaging methods [3].

Besides the clinical appearance, laboratory values seem to be of secondary importance [12]. However, after a closer look, we noticed that the higher the CRP concentration, the more compartments of the hand were infected (see Fig. 3A). However, there was no link between the WBC count and the severity of the hand infection (Fig. 3B). Because of the deep location of closed space accumulations, the typical clinical signs of infection often are not visible, which is a problem in making a proper decision on the surgical strategy [14,15].

The most common signs and symptoms of hand infections are throbbing pain, edema, and restricted finger motion. Specific clinical signs such as Kanavel's signs (septic tenosynovitis) may be obtained in cases of infections in closed spaces. If there is uncertainty about the extent of the infection, measurement of CRP may indicate surgery, as CRP increases dramatically in response to injury, infection, and inflammation [16]. Rothe et al. described that along with the clinical findings, the CRP value is a useful indicator of the treatment efficacy of antibiotics in hand infections [11]. To our knowledge, there are no reports about the value of CRP and WBC counts correlating with the involvement of deep compartments in hand infections. In our study, it seemed that CRP was a valuable marker for the assessment of how many compartments could be affected. Consequently, by the value of CRP, the severity of the hand infection can be estimated well. Thus, the decision to perform surgery can be made easier with a corresponding increase in CRP concentration.

Mixed infections with as many as four pathogens affected significantly more compartments per identified species (Fig. 4). Such microbiologic findings are common in infections caused by bite wounds or in grossly contaminated wounds, in patients with diabetes mellitus, and in infections associated with intravenous drug use [10]. Infection with anaerobic organisms is seen more frequently with diabetes mellitus and intravenous drug use [10]. Bites are a common mode of inoculation [11]. In human bite infection, the usual organisms are Staphylococcus or Streptococcus spp. and gram-negative organisms. Eikenella corrodens is another organism typically associated with human bites [11]. Animal bites are most commonly dog and cat bites with the former far outnumbering the latter [17,18]. These infections tend to arise from Pasteurella, Streptococcus, Staphylococcus, and Bacteroides species [11]. In cases of multiple bacterial infections, the severity may be attributable to lack of broad-spectrum antibiotic therapy at the beginning of the infection or at initial external treatment. Regarding expected mixed infections of the hand, broad-spectrum antibiotic therapy is needed until specific pathogens are identified [1].

Infection of the index and small fingers affected the greatest number of compartments (see Fig. 5). Patel et al. documented that because of a rapid pressure increase in finger infections, the blood supply of the tissues is compromised [3]. This hypoperfusion promotes the transmission of infections in the fingers. In addition, infections can spread quickly through tendon sheaths. This might well be attributable to the high pressure or to the bradytrophic characteristics of the tendon sheaths [3].

In the advanced stages, the infection spreads directly into the surrounding spaces through necrotic and liquefied tissues. Synovial sheaths act as a pathway, which allows the infection to spread. In fingers, the anatomic structures are close to each other, so that local injuries can lead quickly to an effect on the tendon sheaths [3]. In our study, the risk of an initial involvement of the tendon sheath was highest for wounds of the palm of the hand closely followed by wounds of the long fingers. The flexor tendons receive nutrients from a direct vascular supply along the vincular system and through diffusion from the synovial fluid. When infection occurs in the closed system of the flexor tendon sheath, the pressure increases, leading to obstruction of the arterial blood supply ofthe tendons and transmission of the infection into deeper spaces [19]. The deep spaces of the hand include the thenar, midpalmar, Parona, and interdigital web spaces. Deep space infections arise from penetrating inoculation or contiguous spread and, rarely, hematogenous seeding [3]. Therefore infections of these compartments have a high risk of leading to functional loss of the hand [3]. The therapy of these severe hand infections, which balances between radical removal of necrotic infected tissue and preservation of functional structures, requires a great deal of experience from the hand surgeon.

As anticipated, the wrist was highly involved in cases of initial infections of the thumb. In the present study, wrist infections also were caused by long finger infections (see Fig. 6A). This might be explained by fast spread of infection into the midpalmar space and from the index finger into the thenar space, from which infection can spread easily into the wrist (Fig. 6). Contrary to other studies [3,14,15], in this series the spread from wounds of the long-fingers, even the index finger, showed a high risk of involving more compartments and leading to severe infections. Anatomic variations with connection of the tendon sheaths of the index, middle, and/or ring fingers with the radial or ulnar bursa occur in approximately 15% of patients [20]. These variations could lead to different patterns of infection transmission caused by injuries of long fingers. The “typical” horseshoe abscess was observed potentially in two cases (0.5%), which shows the rarity of this hand infection for our population [9]. It is described in the literature that radial and ulnar bursae communicate in 33%–100% of cases [3,20]. Thus, the rarity of horseshoe abscess in this study could be explained by the rare occurrence of the communication between the ulnar and radial bursae in the patient population combined with the early treatment start.

Although early hand infections often can be treated conservatively using antibiotics, early surgical treatment is the safest option [10]. It is necessary because the triggering agent, such as a foreign body, has to be removed in order to enable healing. Surgical treatment should maintain a balance between radical debridement and saving functional structures as far as possible [2]. The skin incision site must be selected so that the entire area of the infection can be accessed and the resulting scars do not lead to any functional disorders. Diagnostics and therapy of hand infections require a detailed knowledge of the anatomy of tendon sheaths, adjacent anatomic compartments, and deep spaces [10]. Even minor injuries of the fingers can spread along the tendon sheath and cause severe hand infections rapidly. During the initial presentation, early operative treatment should be preferred in patients with co-morbidities and advanced age, fingers as entry site, and a CRP concentration >16.1 mg/L (Fig. 7). With higher CRP concentrations, adjacent compartments should be checked. The experienced hand surgeon, particularly in the case of patients with severe hand infection, often indicates a follow-up intervention (“second look”) to guarantee complete infection remediation [21]. In the decision-making process, whether a second-look revision is necessary noticing the number of microbial species and CRP concentration in addition to clinical findings can be helpful (Fig. 7).

OPEN IN VIEWER

FIG. 7.

Decision-making aid in addition to clinical findings of local hand infections.

Because of the retrospective single-center design of this study, caution is required in interpreting our results. Further prospective studies are necessary to validate our findings.

Conclusion

A detailed medical history of co-morbidities, age, cause of the infection, and time between the onset of symptoms and the first presentation is important for the planning of further diagnostic measures and therapy of hand infections. In particular, indications for surgical treatment in patients with co-morbidities should be more permissive. In addition to known clinical signs and imaging techniques, CRP is a reliable marker for assessing the severity of infection, in particular the number of compartments involved. The severity of the infection correlates with the number of pathogenic agents. Because of the different spread of infections depending on the location of the entry site, finger infections in particular should not be trivialized. Superficial infections that do not require surgery can be treated conservatively. However, they should be assessed closely until the infection disappears completely. In case of deep space infections, high CRP concentrations, positive Kanavel's signs, deep bite injuries, foreign body inclusions, etc., beyond use of antibiotic coverage and immobilization with hand elevation early surgical incision and drainage of all potentially communicating spaces and compartments are mandatory.