Patient Risk Factors Associated With Postoperative Complications After Common Hand Procedures

Abstract

Background:

Carpal tunnel syndrome (CTS), trigger finger (TF), and De Quervain tenosynovitis (DQ) are 3 common pathologies of the hand often treated with relatively simple surgical procedures. However, outcomes from these procedures can be compromised by postoperative complications. The aim of this study was to evaluate the association between diabetes, tobacco use, and obesity and the incidence of postoperative complications.

Methods:

We reviewed 597 patients treated surgically for CTS, TF, or DQ from 2010 to 2015. We used bivariate and multivariate analyses to assess independent associations between diabetes, tobacco use, obesity, and surgical complications and compared the incidences with healthier patients without these comorbidities. We also looked at patients with overlapping diagnoses of these comorbidities.

Results:

Bivariate analysis showed that patients with diabetes and smokers were more likely to have a surgical complication. Multivariate analysis showed diabetes and tobacco use as independent predictors of complications. The disease states or combinations placing patients at the highest risk of a postoperative complication were the diabetic-smoker-obese, diabetic-smoker, diabetic-obese, diabetic, and smoker-obese groups. The diabetic-smoker-obese patient population had a 42.02% predicted rate of postoperative complications.

Conclusions:

Diabetes and tobacco use are independent risk factors for complications after operative treatment of CTS, TF, and DQ. Obesity when coexisting with diabetes mellitus (DM) and/or tobacco use increased the risk of complications. When the 3 patient factors evaluated, DM, obesity, and tobacco use, were present, the rate of complications was 42.02%. Careful assessment and discussion should occur before proceeding with operative treatment for simple hand conditions in patients with the risk factors studied.

Introduction

Carpal tunnel syndrome (CTS), trigger finger (TF), and De Quervain tenosynovitis (DQ) are 3 of the most common conditions treated by hand surgeons.^1,2 Surgical treatment of these 3 conditions accounts for more than 500 000 cases per year in the United States ³ and is generally considered to be a safe procedure. ⁴ There is, however, a small subset of patients who experience complications after surgical treatment, including wound complications, surgical site infections (SSIs), and complex regional pain syndrome (CRPS). Although rare, ⁵ these complications result in increased cost to the patient, the surgeon, and the health care system, ⁶ leading to more follow-up visits, increased antibiotic use, and unplanned return to the operating room.

There remains a paucity of literature on patient-centered risk factors for surgical complications after these commonly performed hand surgeries. Three comorbidities that are thought to increase the risk of postoperative complications are diabetes mellitus (DM), ⁷ tobacco use, ⁸ and obesity. ⁹ The association of these factors with complications after common elective surgeries of the hand has not been well studied. The purpose of our study was to assess patient comorbidities and their association with complications after carpal tunnel release (CTR), trigger finger release (TFR), and De Quervain release (DQR). We hypothesized that a history of DM, smoking, and obesity were independent risk factors for complications after these procedures.

Materials and Methods

Study Design

Institutional review board approval was obtained for this study. All procedures were performed by 1 surgeon at a level I trauma center between January 2010 and December 2015. According to institutional policy, all patients received antibiotics before surgical incision. Operative reports were reviewed for any concomitant surgical procedures. Cases involving concomitant CTR, DQR, or TFR were included for analysis. If a staged bilateral CTR was performed, these data were included as 2 patients, as each surgery was considered a separate event. Cases involving concomitant procedures other than CTR, DQR, or TFR; incomplete patient charts; and patients without documented follow-up were excluded.

The electronic medical records (EMRs) were reviewed for demographic information (age and sex), medical history, and medication lists. Patients with diabetes were identified using the International Classification of Diseases, Tenth Revision diagnostic code (E08-E13) or by the use of antihyperglycemic agents or insulin in patients who did not carry the diagnosis in their EMR. This definition was used for the diagnosis of diabetes as opposed to the standard guidelines (hemoglobin A1c of 6.5 or above, fasting glucose of 126 or above, or oral glucose tolerance test of 200 or above), as these laboratory values were frequently unavailable. Tobacco use was defined as habitual tobacco use at the time of surgery, or a 15-pack-year smoking history before surgery. ¹⁰ The average packs-per-day (PPD) was calculated for each smoker. Obesity was defined as body mass index (BMI) of ≥30 kg/m².

We defined a complication as “an undesirable, unintended, and direct result of an operation affecting the patient which would not have occurred had the operation gone as well as could reasonably be hoped.” ¹¹ Complications were divided into 6 categories: surgical-site-specific, structural, nerve-related, medical, ongoing symptoms, and other (eg, CRPS)^12-17 (Table 1).

Table 1.

Summary of Postoperative Complications Considered in this Study’s Cohort.

Category of complication	Type of complication	Examples
1	Surgical-site-specific	Surgical site infection Wound dehiscence Prolonged wound healing Hypertrophic scarring Scar tenderness and pain Transient or excessive wound edema and erythema Reduced range of motion Revision surgery for surgical site Complication
2	Structural	Bowstringing Synovial fistula Finger or joint stiffness
3	Nerve-related	Peripheral nerve injury Digital nerve injury Neuritis Neurapraxia
4	Infectious	Pneumonia Urinary tract infection
5	Ongoing symptoms	Symptom persistence Symptom recurrence Revision surgery for symptom Persistence/recurrence
6	Other	Complex regional pain syndrome

Statistical Analysis

Analysis was performed in XLSTAT software. Descriptive statistics were performed for patient characteristics. Categorical variables were reported as frequencies, and continuous variables were reported as mean ± SD. Patients without the 3 comorbidities of interest formed the control group. Continuous variables were analyzed using the Student t test. Pearson χ² test, or the Fisher exact test when expected values were less than 5 in any contingency table, was performed for each comorbidity independently. Odds ratios with 95% confidence intervals were reported.

A multivariate logistic regression model was generated to assess predictors of complications, controlling for confounding factors among the included variables. The model was statistically valid only if in initial bivariate testing all expected cell frequencies were greater than 1, and no more than 20% of the expected cell frequencies were less than 5. ¹⁸ This ensures that the correct variables are entered into the logistic regression model, eliminating the potential of identifying spurious associations. For the purpose of the model, obesity status was analyzed as a categorical variable (BMI of ≥30 kg/m²). Significant independent predictors of complications were determined to be those with values of P ≤ .05. The Hosmer-Lemeshow goodness-of-fit test assessed whether there was a statistically significant lack of fit between the model and the source data (where a value of P ≤ .05 would indicate a lack of fit). To account for coexisting conditions of more than one of the comorbidities, subgroup analyses were performed on 8 combinations of coexisting conditions. Probability calculations for postoperative complications per subgrouping were performed using marginal standardization, a method of confounder-adjusted logistic regression. This involves first determining the log-odds and odds and then conversion to probability.

Results

Our retrospective review of the EMR yielded 647 patients who underwent the procedures of interest, of which 597 were unique patients who met the inclusion criteria. The breakdown of the procedures consisted of 384 CTR, 134 TFR, and 31 DQR. The mean age of our patient group was 56.80 ± 12.86 years. There were 178 men (29.82%) and 419 women (70.18%), with a mean BMI of 31.61 ± 7.20 kg/m². There were 144 patients with diabetes (24.12%), 120 smokers (20.10%) with mean usage of 0.66 ± 0.32 PPD, and 334 patients with obesity (55.95%). The mean follow-up was 11.22 months (median: 1.96 months).

Complications occurred in 97 patients (16.25%; Table 2). Diabetic status (P = .0001) and smoking history (P = .0015) were significant independent risk factors for complications, whereas obesity was not (P = .132) (Table 3). Patient age (P = .261) and BMI (P = .176) did not demonstrate an association with complication risk.

Table 2.

Number of Complications Occurring per Category Considered.

Category of complication	Type of complication	Number
1	Surgical-site-specific	27
2	Structural	10
3	Nerve-related	7
4	Infectious	0
5	Ongoing symptoms	45
6	Other, complex regional pain syndrome	8

Table 3.

Odds Ratio of Complications per Disease Group.

Group	No. (%)	Total complications (rate, %)	OR (95% CI)	P value
Diabetic	144 (24.12)	38 (26.39)	2.44 (1.54-3.87)	.0001
Smokers	120 (20.10)	31 (25.82)	2.17 (1.34-3.52)	.0015
Obese	334 (55.95)	61 (18.26)	1.41 (0.90-2.21)	.132

Note. OR = odds ratio; CI = confidence interval.

The subgroup analysis of coexisting conditions demonstrated that the populations with the highest predicted percent risk of complications were the diabetic-smoker-obese (42.02) and diabetic-smoker groups (38.69), whereas the obese (11.37) and healthy control (10.05) groups had the lowest predicted risks (Table 4).

Table 4.

Predicted Risk Percentages of Coexisting Conditions.

Disease group D/S/O	n	Complications	Observed risk, %	Predicted risk, %
Healthy −/−/−	191	19	9.947643979	10.0459
Diabetic +/−/−	24	4	16.66666667	21.44935
Diabetic-obese +/−/+	95	24	25.26315789	23.87375
Diabetic-smoker +/+/−	10	4	40	38.68999
Diabetic-smoker-obese +/+/+	15	6	40	42.02063
Smoker −/+/−	38	9	23.68421053	20.51442
Smoker-obese −/+/+	57	12	21.05263158	22.8639
Obese −/−/+	167	19	11.37724551	11.36792

Multivariate analysis (Table 5) showed diabetes (P = .000; odds ratio [OR], 2.45; 95% confidence interval [CI], 1.51-3.96) and tobacco use (P = .001; OR, 2.31; 95% CI, 1.41-3.79) to be statistically significant predictors of a complication. The Hosmer-Lemeshow statistic demonstrated a P value of 1.000, indicating a good fit between the model and the source data.

Table 5.

Multivariate Analysis of Risk Factors.

Variable	Coefficient value (B0, B1, B2, B3)	SE	Pr > χ2	OR (95% CI)
Intercept	−2.192	0.206	<0.0001
Diabetic	0.894	0.245	0.000	2.445 (1.511-3.956)
Smoker	0.838	0.252	0.001	2.311 (1.409-3.790)
Obese	0.138	0.240	0.563	1.148 (0.718-1.837)

Note. OR = odds ratio; CI = confidence interval.

Discussion

This study demonstrated that diabetes and smoking are independent predictors of postoperative complications after 3 common hand procedures. Surprisingly, obesity was not associated with postsurgical complications. Other studies assessing the outcomes of hand surgery have also identified these 2 comorbidities as risk factors for complications.^19,20 Lipira et al ¹⁹ assessed more than 450 000 Medicare patients and found that diabetes, tobacco use, and increased BMI were associated with postoperative infection after open CTR. Similarly, Werner et al ²¹ identified diabetes, smoking, and obesity as risk factors for complications after TFR in the Medicare population.

A myriad of physiological factors may explain these comorbidities as contributory to wound complications. Diabetes has been associated with multifactorial causes of impaired wound healing. ⁷ Tobacco smoke, which contains compounds such as nicotine, carbon monoxide, and hydrogen cyanide, has been shown to cause vasoconstriction, hypoperfusion, and hypoxia, ⁸ also causing improper wound healing. On the contrary, BMI may exert a dose-dependent effect, with BMI >45 kg/m² leading to greater risk of complications following surgeries to the elbow, forearm, and hand. ²² Obesity is often coexistent with other medical comorbidities. In this study, patients who were diabetic, smoker, and obese demonstrated a 42% risk of complications.

Sharma et al ²³ were consistent with our findings, in that diabetes and smoking were included as independent variables in their prognostic risk-adjusted scoring system to predict infection after distal upper extremity procedures, whereas obesity was not. While smoking and diabetes result in systemic changes that negatively affect wound healing at various surgical sites, the soft tissue burden associated with obesity appears to have less impact on surgery at the more distal appendicular sites such as the hand. London et al ²² assessed that the impact of obesity on elbow, forearm, and hand surgeries determined that BMI had a “dose-dependent effect”; when it was greater than 45, patients were at a heightened risk of complication. It is possible that our study may have also concluded this finding if it had been designed to use a cutoff BMI of 45 instead of 30.

The impact of diabetes and tobacco use, as well as obesity, on risk of complications can be diminished if further research can elucidate the appropriate modulations. Long-term glycemic control with an HbA1c < 7 is needed to decrease postoperative infections, ²⁴ but more recent studies have highlighted the importance of perioperative glycemic control, seeing as both diabetes and hyperglycemia in patients without diabetes have been identified as risk factors for SSIs after acute orthopedic intervention and spinal surgeries.^25-27 Despite these findings, there is no established consensus on glycemic control for elective orthopedic procedures, highlighting the need for randomized controlled trials. Regarding smoking cessation, a 2010 review identified a statistically significant reduction in wound healing complications after preoperative smoking cessation ²⁸ and purported that a shorter duration (only 4 weeks) of cessation was needed compared with the previously recommended 6 to 8 weeks. ²⁹ Future studies are needed in the area of hand surgery using this shorter duration of cessation to assess for a decrease in complications. Finally, preoperative weight loss has been shown to paradoxically increase the risk of SSIs in patients undergoing total knee arthroplasty. ³⁰ This increase is thought to have been due to malnutrition, which itself has been identified as an independent risk factor for SSIs. ³¹ There is a need for further research into the role of malnutrition serologic markers such as vitamin deficiencies, total lymphocyte count, albumin, prealbumin, and transferrin in obese patients with complications after elective hand procedures.

Our study had some limitations. As a retrospective study, there are limitations relying on accuracy of documentation and data collection. We minimized this effect by including only patients with charts including all data necessary for the study, including clear documentation of assessment for complications. We also did not have a specific time frame to identify postoperative complications and only followed complications up to their last follow-up appointment. This is an inherent drawback to our retrospective study, and this data collection may invalidate postoperative neurologic symptoms, especially if severe preoperative electromyographic findings were observed. In addition, all procedures were performed at a single academic center by 1 surgeon, potentially giving rise to selection bias in patients treated operatively and affect the generalizability of our results. However, our cohort demographics were similar to those of published literature for patients with CTS, TF, and DQ; 73% of our patients were between the ages of 40 and 70 years, and 68.3% were woman. ³² Finally, a third limitation of our study was that patients with DM were identified as those who had the diagnosis listed in their record or prescribed medications for DM, as pertinent laboratory values were often unavailable on chart review. This has likely resulted in a lower presence of DM within our study cohort.

The results of our study will help hand surgeons identify high-risk patients and adequately counsel them regarding postoperative complications after CTR, DQR, and TFR. Diabetes and smoking were identified by multivariate analysis to be independent risk factors for complications after these procedures. Then, we were able to derive a formula from the regression coefficients to predict the percentage risk of complication, depending on how many of the 3 comorbidities considered a patient had. The patient population with the highest risk of complication (42%) had all 3 comorbidities: diabetes, smoking, and obesity. After determining a patient’s risk for complication, surgeons can appropriately assess whether the potential benefits of operative treatment outweigh the predicted risk of complication, counsel the patients about disease trajectory, and consider closer surveillance and more aggressive intervention for higher risk patients. While perioperative hyperglycemia, smoking status, and malnutrition have been identified as important factors affecting postoperative complication rates, it is clear that future studies are needed to determine the most effective treatment modulations.