Incidence and Determinants of Carpal Tunnel Decompression Surgery in Type 2 Diabetes: The Fremantle Diabetes Study

The Fremantle Diabetes Study (FDS) was a longitudinal observational study in a community of 120,097 people (8). Of 2,258 diabetic patients identified between 1993 and 1996, 1,426 (63%) were recruited and 1,294 had type 2 diabetes. The FDS protocol was approved by the Fremantle Hospital Human Rights Committee. All subjects gave informed consent before participation. Baseline and annual reviews comprised a detailed questionnaire, physical examination, and biochemical tests on fasting blood/urine samples (8). Complications were identified using standard criteria (9).

All deaths and public/private hospitalizations in Western Australia are recorded in the Western Australia Data Linkage System (WADLS) (10). The Confidentiality of Health Information Committee approved linkage of the WADLS and FDS databases. All hospitalizations for open/endoscopic CTD from 1993 until the end of June 2006 were identified using ICD 9-CM codes 04.43/04.45 and 10-AM codes 39331-01/39331-00 for type 2 FDS patients and the Western Australian population. Correct coding was confirmed by chart review in a sample of 28 FDS patients. The incidence of first CTD was determined 1) in FDS patients by dividing the number by the duration of follow-up to CTD or death/end (June 2006) and 2) in the Western Australian population from CTD hospitalizations between January 1993 and June 2006 and corresponding annual population figures. Kaplan-Meier plots of CTD-free survival by sex were compared by log-rank test. Cox proportional hazards modeling with forward conditional variable entry (P < 0.05) and removal (P > 0.10) was used to determine independent predictors of first CTD.

There were 1,284 type 2 FDS subjects without a self-reported/WADLS-documented CTD at baseline. A total of 67 patients (5.8%) had a first CTD during 12,109 years (mean 9.4 ± 3.7), an incidence of 5.5 per 1,000 patient-years. Men and women had a similar incidence (5.8 vs. 5.3 per 1,000 patient-years, P = 0.74). A total of 18 patients (26.9%) had at least two CTDs during follow-up, representing re-operation or decompression in the opposite hand. CTD was performed in a private facility in 76.1% of cases.

In the Western Australian population, 32,836 people had a CTD during 25,226,010 person-years, or 1.30 per 1,000 person-years (95% CI 1.29–1.32). The CTD incidence in the FDS subgroup was therefore 4.2 times higher (P < 0.001), but this is a conservative relative rate because the general population incidence does not exclude further CTDs in the same patient.

Baseline univariate determinants of incident first CTD are shown in Table 1. In Cox modeling, independent associates were higher BMI (hazard ratio 1.05 [95% CI 1.01–1.09]), taking lipid-lowering therapy (2.32 [1.30–4.13]), and being married/in a de facto relationship (2.04 [1.11–3.74], P ≤ 0.021).

The crude incidence of first CTD in our type 2 diabetic patients is between 4 and 14 times that in published general population figures (3) and at least 4.2 times that in the Western Australian population. This difference parallels the threefold increased risk of CTS in type 2 diabetes (11,12). Consistent with a general population study that identified obesity as a strong associate of CTD (13), a 1 kg/m² increase in baseline BMI increased the likelihood of CTD in our patients by 5%. Unlike other studies of CTS determinants in general-population (1,2,14,15) and diabetes-specific (4,12,16–19) studies, we did not find that baseline age, sex, diabetes duration, arthritis, microangiopathy, smoking, or thyroid disease predicted CTD.

Our data suggest that sociodemographic factors are important determinants of CTD in type 2 diabetes. The relationship with marital status could reflect a partner's influence on the decision to pursue CTD. In addition, the ability to pay for a private operation (given long public waiting lists) would be greater with multiple income sources, consistent with the univariate inverse association between household income and CTD in our data. However, the apparent importance of sociodemographic factors compared with recognized CTS associates in diabetes (4,12,16–19) might also indicate that variables such as older age and vascular complications are contraindications to surgery.

Since indications for lipid-lowering therapy in diabetes were less inclusive during FDS recruitment than afterward (20), baseline use of these agents could have been a surrogate for general treatment-seeking behavior including CTD. Uptake of lipid-lowering therapy increased fourfold during the FDS, and there are case reports linking these drugs to neuropathy (21). However, such reports are rare and do not include CTS. Hypercholesterolemia may increase CTS risk (22), and it is possible that dyslipidemic patients were commenced on lipid-lowering therapy after experiencing significant cholesterol-induced median nerve enlargement (22) with a subsequent paradoxical association between statin/fibrate use and CTD incidence.

Our study had limitations. We had no data on the prevalence and incidence of CTS, indication(s) for operation, or relative frequency of clinical review that would have facilitated a more detailed analysis of the high CTD incidence compared with that in the general population. Nevertheless, our findings reflect the relative CTS prevalence in diabetes found previously (1–3). Strengths of the WADLS include its low coding error rate (23) and stable population base (24).

Because of its high prevalence (4), we recommend that CTS assessment for symptoms and signs of median nerve dysfunction should be part of regular complications screening in all type 2 diabetic patients. The present data suggest that many patients with CTS will require CTD for symptom relief.

FDS was funded by the Raine Foundation, University of Western Australia.

We thank FDS patients and staff for support.