High HbA_1c Levels Are Associated With Development of Trigger Finger in Type 1 and Type 2 Diabetes: An Observational Register-Based Study From Sweden

Trigger finger (TF) is a common hand disorder affecting both the flexor tendons in the hand, but also the tight pulley system surrounding the flexor tendons, causing the fingers to lock in flexion. Diabetes is a known risk factor for development of TF, and several studies have linked TF to the more complex syndrome of “the diabetic hand,” together with diagnoses such as carpal tunnel syndrome, ulnar nerve entrapment, and Dupuytren’s disease (1–4). Treatment of TF ranges from conservative methods, such as splinting, occupational therapy (5), and corticosteroid injections, to, if these methods fail, surgery and open release of the A1 pulley (6). However, the best treatment for TF in individuals with diabetes is still under debate (7), and furthermore, studies have reported a higher complication rate after surgery among individuals with TF and concomitant diabetes compared with individuals without diabetes (8). Thus, preventive measurements for TF among the population with diabetes are of importance to avoid both costly surgery and complications.

In a previous American study, investigators found an association between high HbA_1c and TF, although this study did not include stratification for type of diabetes, i.e., type 1 diabetes (T1D) and type 2 diabetes (T2D) (9). Furthermore, poor glycemic control has been linked to both compression neuropathy (10) and autonomic neuropathy (11). Thus, whether strict glycemic control effectively lowers risk of TF among people with T1D or T2D is yet to be determined. Consequently, the aim of this study was to examine whether high HbA_1c was associated with increased risk of TF among individuals with T1D and T2D with use of data from 95,000 individuals in the Swedish National Diabetes Register (NDR).

In this study, data from 2004 to 2019 from the Skåne Healthcare Register (SHR) of the region of Skåne in southern Sweden was cross-linked with data from the NDR. The Region of Skåne is the most southern part of Sweden with ∼1.1 million inhabitants aged ≥18 years. There are public and private caregivers in both primary care and specialized care, and all caregivers are required to transfer data on physician-made diagnoses to SHR (12).

The NDR covers ∼90% of all individuals with diabetes in Sweden and in the Region of Skåne. This study includes individuals with diabetes from 2004 to 2019. The NDR includes individual data on diabetes type, duration of diabetes, medications, laboratory values such as HbA1_c and blood lipid levels, blood pressure examinations, BMI, and complications, e.g., retinopathy and albuminuria. There are several previous studies with descriptions of NDR, used data sources, and definitions of both T1D and T2D (13,14). Individuals with diabetes have their laboratory and anthropometric values, blood pressure, and BMI measured at least once per year, most often in primary care by a registered nurse, where after results are reported back to NDR.

Data on incident TF during 2004–2019 were collected from the SHR and cross-linked to the data from NDR using each individual’s personal identification number, i.e., a 10-digit number unique for each Swedish citizen. The SHR classifies diagnoses according to ICD-10, and for TF the code M65.3 was used (12). Data from both primary and secondary care were included in this study, and only first-time diagnoses of TF during the study period were considered. The diagnosis of TF had to be established after or during the same year as the individual’s diagnosis of diabetes.

HbA_1c was reported in millimmoles per mole and as an individual’s mean value of all registered HbA_1c values during the study period. This method was chosen because glycemic control might vary from year to year and a mean value over several years reflects an individual’s long-term glycemic control better. Smoking was reported as current smoking at an individual’s last registered year or nonsmoking. Systolic blood pressure (mmHg) and BMI (calculated as weight in kilograms divided by the square of height in meters) were also reported as an individual’s mean value for all registered years in the NDR. Finally, for calculation of each individual’s duration of diabetes, the date from the first diagnosis of diabetes to the last registered year in NDR was used. For age calculation, the age of each individual at the last registration year in the NDR during the study period was used.

Quantitative data from NDR, stratified for individuals with and without TF during the study period, are expressed as mean ± SD or median with interquartile range (IQR) when the data are normally or skewed distributed, respectively. Student t test was used for normally distributed continuous data, and Mann-Whitney U test was used for skewed data in comparing group mean and median. For dichotomous variables, a χ² test was used for group comparison.

Individuals were divided into tertiles according to mean HbA_1c with cutoff lines for optimal control ≤48 mmol/mol, acceptable control 48.1–64 mmol/mol, and poor control >64 mmol/mol—well in line with, i.e., an adaptation of, the National Institute for Health and Care Excellence guidelines and the American Diabetes Association guidelines for glycemic targets (15,16). The tertiles were thereafter analyzed in the logistic regression models with the lowest tertile as reference. Finally, when used as a continuous variable, levels of HbA_1c were transformed with z score standardization ([variable level − mean] / SD) before analysis in the regression models. This results in a variable with a mean of 0 and SD of 1, and the data were expressed as odds ratios (ORs) per 1-SD increase in HbA_1c.

For analysis of the effect of high HbA_1c on risk of TF, sex-stratified multivariable binary logistic regression models were created. Two models were used, where the first model included only age at entry in the NDR and HbA_1c. The second model included further adjustment for duration of diabetes, BMI, smoking, and systolic blood pressure, i.e., variables that have been associated with TF and related hand diagnoses in previous studies (7,17–19). Finally, there were two sensitivity analyses performed. The first excluded all individuals ≤40 years at their last registration in NDR, thus excluding younger patients, and the second included only individuals with duration of diabetes >10 years.

All calculations and regression models were performed with stratification by sex and for T1D and T2D separately. We considered a two-sided P value <0.05 as significant, and for the data analysis IBM SPSS statistics for MAC, version 27 (SPSS Inc., Chicago, IL), was used.

The study conforms with the principles of the Declaration of Helsinki. The study was approved by the Swedish Ethical Review Authority (no. 2019-02042).

The data can be applied for by researchers by contacting the KVB committee in Region of Skåne (www.skane.se/en) after an ethics approval from the Swedish Ethical Review Authority.

In total, 9,682 individuals (4,304 women and 5,378 men) with T1D and 85,755 individuals (36,264 women and 49,491 men) with T2D aged ≥18 years were included from the NDR during 2004–2019. There were 486 women and 271 men with T1D and 1,143 women and 1,009 men with T2D who were diagnosed with TF during the study period (Table 1).

Mean HbA_1c was higher for both men and women who were diagnosed with TF during the study period, among both individuals with T1D (P < 0.01) and individuals with T2D (P < 0.01). Furthermore, both men and women who were diagnosed with TF had a longer duration of diabetes, in the case of both T1D (P < 0.01) and T2D (P < 0.01). All quantitative variables are displayed in Table 1.

There was a significant association between high HbA_1c and TF in all logistic regression models, both when with use of HbA_1c as a continuous z score variable and in dividing participants into tertiles according to mean HbA_1c, even after adjustment for confounding factors, i.e., age, smoking, duration of diabetes, BMI, and systolic blood pressure in model II (Table 2). Men with T1D with poor glycemic control had the highest odds of TF (OR 5.33 [95% CI 2.14–13.26], P < 0.001) compared with individuals with T1D and optimal glycemic control.

In the first sensitivity analysis, excluding all individuals ≤40 years at the last entry in NDR did not change the results significantly in the regression models. High HbA_1c was still strongly associated with TF, both when used as a continuous variable and when considered divided into tertiles. Likewise, with only inclusion of data from individuals with duration of diabetes >10 years, high HbA_1c was still highly associated with TF, both when used as a continuous variable and with division into tertiles (Supplementary Tables 1 and 2). Further adjustment in model II did not alter the results significantly (data not shown).

In this study, we present evidence that poor glycemic control among individuals with T1D and T2D is associated with development of TF during a study time of 15 years, using high-quality data from Swedish registers. A dose-response relationship was found with increasing risk of TF with increasing HbA_1c among individuals with diabetes, also after adjustment for age, duration of diabetes, BMI, smoking, and blood pressure. The results are in line with the results of our previous studies on the diabetic hand (4) but also with those of several other studies, indicating that TF is more prevalent among the population with diabetes (2,3). With this study, however, we add individual, long-term data on HbA_1c and glycemic control within the population with diabetes indicating an increased risk for individuals with poor blood glucose control. These results emphasize the importance of primary prevention of T2D, but also the importance of optimal glycemic control among patients with both T1D and T2D in order to prevent diabetic hand complications with subsequent risk of loss of hand function. In terms of clinical practice, the findings of this study can hopefully help motivate both heath care professionals and individuals with diabetes to strive for optimal glycemic control, for patients to reach their individual glycemic targets and to avoid musculoskeletal complications related to their disease.

The pathophysiology behind TF is, after of decades of research, still not fully understood (20,21), especially in the presence of diabetes (7). However, there is a wide agreement that the locking phenomenon is caused by an imbalance in size between the flexor tendon and the surrounding pulley system, most often at the level of the A1 pulley (21), causing the affected finger to lock in flexion (22,23). However, whether the underlying cause is primarily located within the flexor tendon or the pulley or both is still under debate (24). There are several imaging studies with ultrasonography (25–27), implying a thickening of the A1 pulley among patients with TF, as well as histopathological biopsies studies, indicating fibrosis (28) and increased collagen deposition (29) in the A1 pulley in patients with TF, with increased pulley stiffness (26) as a possible result. Furthermore, constriction of the A1 pulley in cadaver hands successfully caused a trigger phenomenon in the studied finger (30). At the same time, the literature also reports thickening of the flexor tendon (31,32) with subsequent formation of noduli (24) as a putative cause for the locking phenomena in TF. Finally, biopsies from the tenosynovium in patients with TF showed signs of both edema and inflammation (33), possibly adding volume to the already limited space available for the flexor tendon (23) (Fig. 1). Taking this together, the biomechanical cause of TF is most likely multifactorial, with a combination of tendinopathy and pulley thickening as the causal factor behind the triggering phenomenon.

In our study, we found increasing risk of TF with higher levels of blood glucose, measured as HbA_1c, both in T1D and in T2D. Men with T1D in the highest tertile of blood glucose (mean HbA_1c >64 mmol/mol) were five times as likely to develop TF compared with those in the lowest tertile (mean HbA_1c ≤48 mmol/mol), even after adjustment for age and duration of diabetes. These results are in line with those of several previously published studies, indicating diabetes as a risk factor for TF (1,7,9,34). However, whether duration of diabetes or glycemic control increases the risk is under debate (7). This study provides additional evidence that glycemic control is indeed of importance in development of TF in providing duration-adjusted, large-scale data on glycemic control and independent risk calculations for men and women with T1D and T2D separately.

How hyperglycemia and diabetes affect the pulley system and the biochemical alterations in the tendon is not fully understood, and studies including biopsies on patients with TF and concomitant diabetes are scarce. The few studies with available biopsies from the A1 pulley in patients with TF and concomitant T2D reported signs of increased oxidative stress and high levels of oxidative protein products in the pulley tissue (35), as well as signs of increased neovascularization in the A1 pulley in hyperglycemic patients (36), a possible sign of oxidative stress due to diabetes-induced hypoxia. Furthermore, diabetes also induces tendinopathy with subsequent tendon swelling (37), possibly due to an increased amount of intracellular hyperosmotic proteins (sorbitol) causing edema through the polyol pathway (38). Formation of so-called advanced glycation end products (AGEs), a result from prolonged hyperglycemia, disrupts normal tendon homeostasis, further adding to the tendinopathy (39). Advanced glycation end products have also been associated with both micro- and macrovascular complications of diabetes, as well as with joint stiffness (40). Thus, a combination of increased stiffness of the A1 pulley due to oxidative stress, together with diabetes-induced tendinopathy and tendon swelling, might explain the increased risk of TF among individuals with T1D and T2D with poor blood glucose control (Fig. 1). The dose-response relationship found in this study between levels of HbA_1c and TF further adds to the evidence that hyperglycemia is (one of) the causal explanatory factors behind this association.

There are several strengths to this study, the size of the study population first but also the high-quality standards of the used registers. The NDR covers the vast majority of all patients with diabetes in the Region of Skåne and includes yearly follow-up of both anthropometry and laboratory values. The SHR also has a high coverage rate, since it is mandatory for caregivers in the region to transfer data on physician-made diagnoses to the register. The study size, enabling stratification not only for sex, but also for diabetes type, is one of the major strengths of this study. To our knowledge, this is one of the largest studies of diabetes and the risk of TF, including data on both T1D and T2D.

Nevertheless, there are some limitations to the current study, the first one being its retrospective nature and dependence on accurate registration in the NDR and SHR. There is always a risk of inaccurate registration, in all-quality registers. NDR registrations are made by a registered nurse, often specialized in diabetes care, or by automatic transfer from electronic heath records, whereas the SHR relies on automatic transfer of physician-made diagnoses to the register. In this way, there is no recall bias for any of the individuals in this study. Furthermore, we were only able to include individuals living in the Region of Skåne. Individuals moving to another region were excluded from the year they moved away from the region. To this day, there are no national registers that include data on hand surgical diagnoses from both primary and secondary care in Sweden, and we therefore had to rely on regional data registers. Moreover, there might be other known or unknown confounding factors, for which we were not able to adjust in this study, e.g., alcohol consumption and upper-extremity neuropathy, affecting the results presented. Neuropathy is registered in NDR, but only on the basis of monofilament testing of the feet, and we did not consider this variable reliable. Finally, the study was based on measurements of HbA_1c, reflecting an individual’s glycemic control during the last 3 months (15). Most individuals in NDR have HbA_1c reported once per year. Since this date was not available for each individual, we were not able to match it with the corresponding date of the diagnosis of TF. Instead, we calculated the mean value of HbA_1c from all registered years in NDR, which better reflects the individual long-term glycemic control. Data on glycemic control adjacent to the diagnosis or surgery of TF would make an interesting case for a smaller case-control study.

Hyperglycemia increases the risk of developing TF among individuals with T1D and T2D. Optimal treatment of diabetes seems to be of importance for prevention of diabetic hand complications, such as TF.

Funding. This study was funded by Stig and Ragna Gorthons Foundation, Skåne University Hospital, Lund University, the Swedish Diabetes Foundation, the Swedish Research Council (grant 2021-01942), and the Regional Agreement on Medical Training and Clinical Research (ALF) between Region of Skåne and Lund University.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.R. and M.Z. designed the study. M.R. wrote the first draft of the manuscript. All of the authors discussed the data and results and contributed equally to the discussion and content of the manuscript. All of the authors reviewed and accepted the final version of this manuscript before publishing. M.R. and L.B.D. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at IFSSH, IFSHT & FESSH Combined Congress, London, U.K., 6–10 June 2022.