Identification of Clinical and Demographic Predictors for Treatment Modality in Patients With Carpal Tunnel Syndrome

Abstract

Background:

Carpal tunnel syndrome (CTS) is the most common entrapment neuropathy. There are multiple treatment modalities for CTS. This study examines both clinical and demographic predictors for initial treatment modality of CTS.

Methods:

Patients diagnosed with CTS between February 2015 and October 2020 with a hand clinic visit within 6 weeks before treatment were included in our study. Patients completed Patient-Reported Outcomes Measurement Information System Pain Interference, Physical Function, and Depression and had complete data on relevant predictor variables. Primary outcomes were treatment group: (1) injection only; (2) release only; and (3) injection followed by release. Bivariate analyses and multinomial logistic regression was used to identify statistically significant variables and independent predictors associated with the treatment groups, respectively.

Results:

A total of 1409 patients fit our inclusion criteria. Sex, age, body mass index (BMI), race, ethnicity, Pain Interference, and Depression were statistically significant predictors for treatment group in bivariate analysis (P < .05). In multivariable analysis, adults older than 65 years were less likely to receive either injection only or injection followed by release (odds ratio [OR]: 0.56 and OR: 0.52, respectively; P < .01). Overweight (BMI: ≥25) individuals were less likely to receive injection only (OR: 0.45; P < .01). Women were more likely to have either injection only or injection followed by released (OR: 1.50 and 1.55; P < .01). Similarly, black, Indigenous, or Persons of Color had an increased odds of injection only and injection followed by release (OR: 1.61 and OR: 1.69, respectively; P < .05).

Conclusions:

Sex, age, BMI, race, and ethnicity were found to be independent predictors of treatment modality for CTS.

Introduction

Carpal tunnel syndrome (CTS) is the most common entrapment neuropathy and is estimated to affect approximately 3% to 6% of adults. ¹ There are patient factors associated with CTS incidence, such as women being 3 times more likely than men to have CTS ² and comorbidities associated with CTS include diabetes and obesity. ³ In a meta-analysis of 58 studies, Shiri et al. found that each 1-unit increase in body mass index (BMI) increased the risk of CTS by 7.4%. ⁴ In addition, occupation may have implications on developing CTS with potential increased risk in patients that have consistent exposure to repetitive movements and/or vibratory tools. ⁵ Race has also been studied in terms of CTS, with Garland et al reporting CTS rates 2 to 3 times higher in white individuals as compared with black individuals in US Navy personnel. ⁶ Carpal tunnel syndrome also appears to be rare in certain populations as Goga found that patients who are black and from South Africa showed very low incidence of CTS. ⁷

There are multiple treatment modalities for patients with CTS. These modalities range from surgical decompression to conservative treatments of splinting, corticosteroids, physical therapy, and therapeutic ultrasound. ⁸ The mechanism of the use of corticosteroids in the treatment of CTS is to reduce the swelling within the carpal tunnel. Steroid injection has shown to be effective in treatment of CTS in multiple studies.^9-11 Green et al reported 81% of patients had complete relief that lasted from 1 day to 45 months, with the majority developing recurrent symptoms between 2 and 4 months. ¹⁰ In addition, 35% of the patients who received injections eventually underwent carpal tunnel release. In 2007, a Cochrane review concluded that steroid injection for CTS provides significant clinical improvement compared with placebo 1 month after injection. ¹² Surgery is not without risk as shown by Kaile and Bland et al who demonstrated that patients who underwent carpal tunnel release have a greater risk of complications when compared with steroid injection, which has a serious complication rate of <0.1%. ¹³

With these data and the known risks of surgery, is it possible to predict which patients should receive injection or release for CTS? To answer this question, we need to first understand which patient factors are currently predictors for CTS treatment modality.

Patient-reported outcomes are an important part of care and becoming increasingly used. As such, we also looked at our Patient-Reported Outcomes Measurement Information System (PROMIS) data to determine whether these preoperative scores could predict treatment type. Patient-Reported Outcomes Measurement Information System is a Patient-Reported Outcome Measures (PROMs) that are scored in a standardized manner (ie, matched to the average person for the 2000 US Census) with a mean t score of 50 and standard deviation of 10. ¹⁴ Higher PROMIS scores indicate more of the construct being measured. As such, PROMIS Physical Function (PF) t scores represent better physical function. In addition, higher PROMIS Pain Interference (PI) t scores signify more pain interference and higher PROMIS depression scores represent worse depressive symptoms. ¹⁵

The goal of our study is to evaluate the clinical and demographic predictors for initial treatment modality (injection only, release only, or injection followed by release) for CTS.

Materials and Methods

We conducted a retrospective cohort study of patients diagnosed with CTS between February 2015 and October 2020 at a single academic medical center.

Eligible patients were identified from a patient-reported outcomes database. This database included numerous patient characteristics and PROMIS scores. Inclusion criteria are a hand clinic visit within 6 weeks before treatment for CTS. The reason for this time restriction was to maximize the likelihood that the PROMIS scores are reflective of the CTS symptoms and are not due to other conditions that might be present with greater time and confound our analyses. Diagnosis of CTS was based on history and clinical exam, with CTS-6 scores or electrodiagnostic studies. Our criteria for CTS-6 scores were 12 or greater as this indicates greater than 80% probability that the patient has CTS.

Treatment included carpal tunnel release only, injection only, or injection followed by carpal tunnel release. In addition, patients had to complete PROMIS PI, PF (V 1.2, V 2.0), and Depression Computer Adaptive Test, and having complete data on relevant predictor variables.

Data from the patient-reported outcomes measure database were merged with demographic and clinical data from the electronic medical record. The demographic and clinical data included age, BMI, sex, race, ethnicity, marital status, and smoking status. The patient-reported outcomes measure database included PROMIS PI, PF, and Depression, which are administered at our institution as part of routine clinical care.

The primary outcome was defined as treatment group (1) injection only; (2) release only; and (3) injection followed by release. The independent predictor variables included age (<65 or ≥65 years), sex (male or female), race (white or black, Indigenous, or Persons of Color [BIPOC]), ethnicity (Hispanic or non-Hispanic), BMI (normal or >normal [>25]), marital status (married or other [divorce, widowed, separated, etc]), insurance (Medicaid and self-insurance or other [commercial/private, Medicare, motor vehicle, workers compensation]), smoking status (active or nonactive), PROMIS PI (<60 or ≥60), PROMIS PF (<40 or ≥40), and PROMIS Depression (<60 or ≥60).

Bivariate analyses were used to identify variables statistically significantly associated with the treatment groups. These variables were then entered into a multinomial logistic regression model to identify independent predictors of treatment groups. In the multinomial logistic regression, the release-only group was used as the reference group, to which injection only and injection, followed by release group were compared. Odds ratios (ORs), 95% confidence intervals (CIs), and P values were calculated to quantify the relative odds of injection only and release only compared with injection followed by release.

Results

Our analysis included 1409 patients. Most of the patients in our study were under the age of 65 years (965; 68%), women (871; 62%), and white (1225; 87%) (Table 1). In terms of primary outcome, majority of patients received release only: 880 (62%), followed by injection only: 281 (20%) and finally, injection followed by release: 248 (18%).

Table 1.

Patient Characteristics (n = 1409).

Characteristic	Value
Age, y, n (%)
<65	965 (68)
≥65	444 (32)
Sex, n (%)
Male	538 (38)
Female	871 (62)
BMI, n (%)
<25	216 (15)
≥25	1193 (85)
Race, n (%)
White	1225 (87)
BIPOC	184 (13)
Ethnicity, n (%)
Hispanic	78 (6)
Not Hispanic	1331(94)
Married, n (%)
No	592 (42)
Yes	817 (58)
Insurance, n (%)
Medicaid or self-pay	229 (16)
Other insurance	1180 (84)
Active smokers, n (%)
Active	224 (16)
Not active	1185 (84)
PROMIS Pain Interference
<60	768 (55)
≥60	641 (45)
PROMIS Physical Function
>40	486 (34)
≤40	923 (66)
PROMIS Depression
<60	1196 (85)
≥60	213 (15)

Note. BMI = body mass index; BIPOC = black, Indigenous, or Persons of Color; PROMIS = Patient-Reported Outcomes Measurement Information System.

In bivariate analysis, age (<65 or ≥65) (P < .01), sex (male or female) (P < .01), BMI (normal or >normal [>25]) (P < .01), race (white or BIPOC) (P < .01), ethnicity (Hispanic or non-Hispanic) (P < .01), PROMIS PI (<60 or ≥60) (P = .04), and PROMIS Depression (<60 or ≥60) (P = .05) were found to be statistically significant predictors for treatment groups (Table 2). There was no significant difference in treatment modality for CTS (release only, injection only, and injection than release) between any of the following bivariate analysis: marital status (married or other [divorce, widowed, separated, etc]) (P = .11), insurance (Medicaid and self-insurance or other [commercial/private, Medicare, motor vehicle, workers compensation]) (P = .51), smoking status (active or nonactive) (P = .78), and PROMIS PF (<40 or ≥40) (P = .30) (Table 2).

Table 2.

Clinical and Demographic Factors Associated with Treatment Modality in Bivariate Analysis.

Demographics	Injection only N = 281 N (%)	Release only N = 880 N (%)	Injection then release N = 248 N (%)	P value
Age, y				<.01
<65	216 (76.87)	554 (62.95)	195 (78.63)
≥65	65 (23.13)	326 (37.05)	53 (21.37)
Sex				<.01
Female	197 (70.11)	500 (56.82)	174 (70.16)
Male	84 (29.89)	380 (43.18)	74 (29.84)
BMI				<.01
<25	65 (23.13)	115 (13.07)	36 (14.52)
≥25	216 (76.87)	765 (86.93)	212 (85.48)
Race				<.01
BIPOC	52 (18.51)	87 (9.89	45 (18.15)
White	229 (81.49)	793 (90.11)	203 (81.85)
Ethnicity				<.01
Hispanic	30 (10.68)	32 (3.64)	16 (6.45)
Not Hispanic	251 (89.32)	848 (96.36)	232 (93.55)
Married				.11
Married	153 (54.45)	507 (57.61)	157 (63.31)
Not married	128 (45.55)	373 (42.39)	91 (36.69)
Insurance				.51
Medicaid or self-pay	52 (18.51)	139 (15.80)	38 (15.32)
Other insurance	229 (81.49)	741 (84.20)	210 (84.68)
Smoking				.78
Active smoker	47 (16.73)	141 (16.02)	36 (14.52)
Not active smoker	234 (83.27)	739 (83.98)	212 (85.48)
PROMIS Pain Interference				.04
<60	140 (49.82)	503 (57.16)	125 (50.40)
≥60	141 (50.18)	377 (42.84)	123 (49.60)
PROMIS Physical Function				.30
>40	173 (61.57)	584 (66.36)	166 (66.94)
≤40	108 (38.43)	296 (33.64)	82 (33.06)
PROMIS Depression				.05
<60	236 (83.99)	761 (86.48)	199 (80.24)
≥60	45 (16.01)	119 (13.52)	49 (19.76)

Note. BMI = body mass index; BIPOC = black, Indigenous, or Persons of Color; PROMIS = Patient-Reported Outcomes Measurement Information System.

In the multinomial logistic regression analysis, the injection-only group and the injection followed by release group were compared with the release-only group. Patients 65 years or older were less likely to receive injection only when compared with release only for CTS (OR: 0.56; 95% CI, 0.41-0.78; P < .01) (Table 3). Female patients, when compared with male patients, were more likely to receive injection only when compared with release only for CTS (OR: 1.50; 95% CI, 1.12-2.03; P < .01). Patients with a BMI of 25 or greater were less likely to receive injection only when compared with release only for CTS (OR: 0.45; 95% CI, 0.32-0.65; P < .01). Black, Indigenous, or Persons of color patients were more likely to have injection only compared with release only for CTS when compared with white patients (OR: 1.61; 95% CI, 1.07-2.42; P < .05). Patients who were of Hispanic ethnicity treated for CTS were more likely to receive injection only compared with release only than patients who are not Hispanic (OR: 2.54; 95% CI, 1.46-4.42; P < .01).

Table 3.

OR of Variable Effect on Predicting Treatment Group in Multivariable Analysis.

Group	Variables	OR	95% CI	Overall type III analysis of effects, P value
Injection only compared with release only	Age, >65 y	0.56	0.41-0.78	<.01
	Sex, female vs male	1.50	1.12-2.03	<.01
	BMI, >25	0.45	0.32-0.65	<.01
	Race, BIPOC vs white	1.61	1.07-2.42	.01
	Ethnicity, Hispanic vs not Hispanic	2.54	1.46-4.42	<.01
Injection followed by release compared with release only	Age, >65 y	0.52	0.37-0.72	<.01
	Sex, female vs male	1.55	1.14-2.12	<.01
	BMI, >25	0.83	0.55-1.26	<.01
	Race, BIPOC vs white	1.69	1.12-2.56	.01
	Ethnicity, Hispanic vs not Hispanic	1.30	0.68-2.50	<.01

Note. OR = odds ratio; CI = confidence interval; BMI = body mass index; BIPOC = black, Indigenous, or Persons of Color.

Similar results were seen in the injection followed by release group, compared with release-only group. Patients 65 years or older were less likely to receive injection followed by release, compared with release only for CTS than patients younger than 65 years of age (OR: 0.52; 95% CI, 0.37-0.72; P < .01) (Table 3). Female patients were more likely to receive injection followed by release, compared with release only for CTS than male patients (OR: 1.55; 95% CI, 1.14-2.12; P < .01). Overweight or obese patients (BMI, >25) were less likely to receive injection followed by release, compared with release only for CTS than normal weight patients (OR: 0.83; 95% CI, 0.55-1.26; P < .01). Black, Indigenous, or Persons of Color patients were more likely to receive injection followed by release, compared with release only for CTS than white patients (OR: 1.69; 95% CI, 1.12-2.56; P = .01). Finally, Hispanic patients were more likely to receive injection followed by release, compared with release only for CTS than non-Hispanic patients (OR: 1.30; 95% CI, 0.68-2.50; P < .01).

Discussion

We sought to evaluate numerous patient characteristics, including PROMIS outcome scores, to predict treatment modality (injection only, release only, or injection followed by release) for CTS. In our analysis, we identified that sex, age, BMI, race, and ethnicity were independent predictors of treatment modality in our multivariable analysis. While PROMIS PI and PROMIS Depression were identified in bivariate analysis, they were not statistically significant in our adjusted analyses.

In our study, we found that overweight or obese individuals were less likely to receive injection only. Shiri et al described that each 1-unit increase in BMI significantly increases the risk for CTS. ⁴ While this study does not specify severity of carpal tunnel for overweight or obese patients, it is possible that with increased risk comes increased severity. However, there was a similar level of overweight or obese patients receiving injection followed by release, compared with release only. This may signify that patients who are overweight or obese may get some relief from injection, but eventually their CTS progresses and ultimately they receive surgical intervention.

Our analysis has shown that females were more likely to have either injection only or injection followed by releases compared with male patients. A possible explanation for these results is that male patients may have more severe nerve impairment when compared with females. This is consistent with what Padua et al reported, where male patients presented with worse neurophysiological findings than female patients. ¹⁶ Our data would suggest that men have increased CTS severity when compared with females. Further research is warranted to use electrodiagnostic studies to confirm this possible explanation.

In addition, treatment differences according to race and ethnicity were observed in our study. Patients who were BIPOC had increased odds of injection only and injection followed by release. In addition, Hispanic patients were more likely to have injection only. While we were unable to find any studies that specifically analyzed race and treatment for CTS, there are multiple possibilities for our findings. One explanation of this difference may be related to access to care in terms of carpal tunnel release surgery. One related example in the field of orthopedics is a study done by Singh et al that showed that black people covered by Medicare were 36% less likely to receive a total knee arthroplasty than white people in 1991. ¹⁷ This gap may be increasing since they further found that in 2008 that disparity was up to 40%. ¹⁷ In addition, Chen et al used US National Inpatient sample (NIS) data from 2002 to 2007 and found that the probability of receiving a total hip arthroplasty or total knee arthroplasty was higher for whites than blacks or Hispanics. ¹⁸ This finding could also be due to potential unconscious or implicit bias. Santry et al explained the role of unconscious bias in surgical safety, which proposed that race may trigger an unconscious bias in medical decision-making. ¹⁹ This bias may lead to unequitable care for minority patients. They recommend trying to acknowledge our own unconscious bias along with spending time to understand patients’ sociocultural background can lead to more equitable care. In addition, a study about racial disparities in communication between orthopedic surgeons and patients by Levinson et al. found that the content of informed decision-making does not differ by race. However, patient ratings on communication were significantly higher for white patients. ²⁰ While the content of informed decision-making is similar, this study may show that BIPOC patients could benefit from improved communication from orthopedic and hand surgeons to enhance cultural communication, build rapport, and show empathy. A systematic review by Maina et al. focused on implicit racial/ethnic bias in health care providers using the implicit association test, which found 6 of the 14 studies that displayed that implicit bas was associated with disparities in treatment recommendations, expectations of therapeutic bonds, pain management, and empathy. ²¹ As well as, they noted that providers with stronger implicit bias demonstrated poorer patient provider communication.

Within our results, we found that PROMIS scores were not independent predictors of treatment modality. This may show that patient-reported outcomes cannot predict the treatment modality for CTS. This is interesting since one would expect that if the patient had worse PROMIS PI, PF, or Upper Extremity (UE), the patient would be more willing to undergo surgery initially for permanent solution to their problem. Perhaps patients would like injection in the office for immediate relief and then evaluate for release if symptoms return.

One possible explanation for patients aged 65 years or older to be less likely to receive either injection only or injection followed by release is that older patient may have had CTS for a longer period and more advanced symptoms. At this point, surgery is the only treatment that may provide a cure. ²² Younger patients may present with less severe disease such as mild or moderate CTS where treatment may be guided by splinting, nonsteroidal anti-inflammatory drugs, activity modifications, and steroid injections. ²² Previous literature aligns with this hypothesis reporting that older patients with more comorbidities are more likely to get release for CTS. ²³ We based our diagnosis on a combination of history, clinical examination, and diagnostic studies, including CTS-6 and electrodiagnostic studies (EDX). Even with these modalities, there may be questions as to whether symptoms are from median nerve compression and the injection can help predict response to surgery. ²⁴ In addition, some patients may be hesitant to proceed with surgery even though it is recommended by the provider and opt for an injection first.

There are multiple limitations within our study. First, our study is from a single institution with multiple providers in the northeast United States. It is possible that at another institution including different geographical location or culture, patients presenting with CTS may be treated with different proportions of injection only, release only, and injection followed by release, than at our institution of injection only (281; 20%), release only (880; 62%), and injection followed by release (248; 18%). The biggest limitation is that we did not evaluate the electrical studies for severity of compression, or the quantitative CTS-6 scores, which may impact the treatment recommendations, as mild changes are often treated with injection and severe changes are often treated with surgery. In addition, we used CTS-6 score of 12 as our diagnostic score and patients with lower than 12 can also have CTS. The use of electrodiagnostic studies may be able to provide additional insight into our findings and we are interested in this as a future area of research. Further studies should evaluate electrodiagnostic results and quantitative CTS-6 scores to evaluate the treatment modality. Using electrodiagnostic results of mild, moderate, and severe as well as score cutoffs for CTS-6 will help elucidate the CTS severity and treatment modality. Continued research is warranted, and this may help elucidate some of the associations that were observed with demographic characteristics such as age, sex, race, and ethnicity. In addition, in the observational assessment of treatment choice, confounding by indication may be present in our data. There may be additional factors that influenced treatment modality that were not evaluated in our analyses.

Conclusion

Within our study, we found sex, age, BMI, race, and ethnicity to be independent predictors of treatment modality for CTS. Women and BIPOC were more likely to receive injection only or injection followed by release, whereas older patients (>65 years old) were less likely to receive either injection only or injection followed by release. Overweight individuals were less likely to receive injection only. Finally, Hispanic patients were more likely to receive injection followed by release for CTS. There are multiple potential explanations for these findings such as increase severity or duration of dealing with CTS symptoms. In addition, the thought of getting relief more quickly with injection may also impact patient decisions.

While there are no actionable findings in our study, we feel it is important for providers to be aware of potential inequities of treatment for all conditions and use our findings to encourage further research.