Clinical and Laboratory Factors Associated With Wound Healing in Patients With Pyoderma Gangrenosum: A Retrospective Study

Pyoderma gangrenosum (PG) is a rare, often chronic, inflammatory, ulcerative condition that often affects the lower extremities. Sterile neutrophilic dermal infiltrate on biopsy and an irregular, undermined border with a purulent base on clinical examination are hallmarks of the condition.¹

The rarity of PG and poor understanding of the condition make it challenging to study outcomes and prognosis. It has been reported that there may be prognostic utility in laboratory markers such as neutrophil-to-lymphocyte ratio, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) level in patients with ulcerative or inflammatory conditions.^2-5 It has also been noted that not only is it challenging to clinically diagnose wound infections, but that chronic wounds tend to exhibit proinflammatory changes in the wound environment, such as infection or cellular senescence, that ultimately promote wound chronicity.^6,7 Immune dysregulation has also been noted to be related to wound size.⁸ Certainly, there is substantial biologic plausibility for a prognostic inflammatory marker in wound healing. Therefore, the authors of this research brief undertook an exploratory analysis of the relationship between readily available clinical factors and target ulcer healing within 6 months from the first date of service (first clinic visit) in patients with ulcerative PG.

A retrospective review of the Oregon Health & Sciences University health record was conducted between January 2015 and December 2018. Two dermatologists independently reviewed charts and confirmed a diagnosis of PG using the diagnostic criteria of Su et al.⁹ In brief review, the diagnostic criteria by Su et al⁹ requires both major criteria of rapidly progressing, painful cutaneous ulcer with undermined, irregular, and violaceous border; infectious and other causes of cutaneous ulcer have been ruled out; and at least 2 of 4 minor criteria. The minor criteria include (1) pathergy or cribriform scarring, (2) PG-associated systemic disease, (3) sterile dermal neutrophilia ± mixed inflammation ± lymphocytic vasculitis, and (4) rapid treatment response to systemic corticosteroids.⁹ Statistical modeling was limited by the size of the data set relative to the number of potential predictor variables. To explore the role of inflammation in wound healing in patients with PG, 2 logistic regression models were fit—one focused on ESR as a primary covariate adjusted for age and hemoglobin level, and the other focused on CRP level adjusted for body mass index and smoking history. These covariates were chosen for their potential associations with ESR and CRP level and the outcome to reduce the risk of confounding.⁴

In total, there were 68 patients available, but some had missing data for certain measurements that led to their exclusion in certain calculations. Sixteen patients had ulcers that healed within 6 months, and 50 ulcers that did not. The CRP level was high (≥10 mg/L) in 20 unhealed ulcers (40%) and 3 healed ulcers (19%). An abnormal ESR (≥20 mm/hour) was noted in 13 healed ulcers (81%) and in 33 unhealed ulcers (66%). The neutrophil-to-lymphocyte ratio was greater than or equal to 4.0 in 7 healed target ulcers (44%) and in 31 unhealed target ulcers (62%). Regarding classically associated immune conditions, 47% of patients (8/17) had an abnormal colonoscopy (inflammatory bowel disease or malignancy), 12% (3/25) had a positive rheumatoid factor test, and 19% (5/26) had a positive antinuclear antibody test (Table 1A, Table 1B, Table 1C).

The odds of wound healing at 6 months for patients with an abnormal CRP level was 0.19 times the odds of healing (95% CI, 0.03–0.77) for those with normal CRP levels (<10 mg/L), after accounting for body mass index and smoking. Similarly, after accounting for age and hemoglobin level, those with high ESR experienced 2.81 times the odds of healing (95% CI, 0.69–14.58) at 6 months compared to patients with normal ESR (Figure).

Table 2 and Table 3 present the results of general association tests for each ESR and CRP level in relation to the available comorbidities. Care should be taken in interpreting the data because of the limited sample size.

After adjustment, increased CRP level was associated with lower odds of healing; however, further research is warranted to investigate the prognostic utility of CRP. The relationship between ESR and odds of healing should be interpreted cautiously because the wide 95% CI for the adjusted odds ratio reflects marked uncertainty in the estimate despite accounting for age and hemoglobin level. The literature suggests the potential for using either ESR or CRP level for monitoring inflammatory states, including autoinflammatory, neutrophilic conditions such as hidradenitis suppurativa, and it indicates that ESR may be the more useful of the 2 markers because it changes more slowly than CRP level.^4,10 It is known, however, that ESR and CRP level are often nonspecific, and these values should be used in conjunction with other clinical information⁴ (Table 2, Table 3).

This exploratory analysis is limited by a retrospective design and small sample size in the setting of a rare, diagnostically challenging disease. Adjusting for additional covariates is particularly challenging from a numerical perspective. For example, few patients had either congestive heart failure (1/68) or chronic kidney disease (3/68), which prevented adjustment based on these covariates. Adjustment for infection was also challenging because many ­patients with PG have bacterial colonization; because infectious signs and symptoms are also signs of PG, it can be challenging to differentiate infection from colonization. As previously noted, diagnosing infections in chronic wounds is challenging, even for physicians with a background in wound care, and excessive inflammation (including resulting from wound infections) can impair the wound healing process.^6,7 For these reasons and to err on the side of caution, all patients who presented to the clinic with signs and symptoms of infection and positive wound culture were treated for at least a wound infection. All patients in this sample presented with chronic wounds, precluding a comparison between acute and chronic wounds; however, investigating inflammatory markers in this setting may be more pertinent for the physician with expertise in chronic wounds because referral patterns are likely directly linked to wound chronicity. Further, treatment of underlying conditions should receive consideration because some of these therapies may affect PG wound healing. A larger study is required to investigate these confounding variables without compromising the statistical reliability of the estimated relationships.

Despite the caveats common to studying rare and diagnostically challenging pathologies, this research brief serves as a basis for further research into inflammatory markers as possible indicators of healing or as prognostic markers for wound healing in patients with PG.