Introduction

The prevalence of diabetes mellitus (DM) is increasing in the UK, where it has been estimated that there are currently 3.8 million adults with DM, equivalent to 8.6% of the adult population [1]. This is an increase from the estimate of 2.7 million (6% of the adult population) in 2013 [2]. Recent studies have shown that 8.8% of surgical referrals to all specialities are for persons with diabetes, and that 10–15% of those who ultimately undergo surgery have DM [3, 4]. This figure is expected to increase alongside increases in the prevalence of DM.

Poor perioperative diabetes control has been shown to be associated with adverse outcomes in almost every surgical speciality [5,6,7,8,9,10,11,12,13,14,15,16,17,18], although it is acknowledged that these data show an association and not causation. However, given the strength of the association, national guidelines recommend that glycemic control be optimized prior to elective surgery—i.e. the patient should achieve a glycated hemoglobin (HbA1c) of ≤ 69 mmol/mol (8.5%), where it is safe to do so [4, 19]. To date there are few data on how closely clinicians adhere to this standard at the preoperative assessment, or whether this target is achievable.

The aim of this single-center study was to determine how many orthopedic and vascular surgery patients with DM in a convenience sample had a preoperative HbA1c above the recommended target of 69 mmol/mol (8.5%), and how many ultimately underwent surgery despite having an HbA1c which exceeded the target.

Methods

We carried out a retrospective review of a convenience sample of 1000 consecutive vascular and orthopedic patients who had a preoperative assessment between 15 January 2016 and 20 February 2017 at the Norfolk and Norwich University Hospital (NNUH), a 1000-bed teaching hospital in the East of England.

All patients who require an elective vascular or orthopedic procedure are seen in a preoperative assessment clinic in the 2 weeks preceding their surgery. This clinic is run by experienced nursing staff with anaesthetic medical support. For those with significant co-morbidities, the assessments are undertaken by the anaesthetists themselves.

Patients attending their assessment were identified using the hospital’s patient administration system (PAS). Details of patients’ preoperative assessment were found using internal electronic databases (Orsos Web Periop [Allscripts, Manchester UK] and Bluespier [Droitwich, UK]). Sunquest ICE (Integrated Clinical Environment, an online pathology requesting and reporting system [Tucson, AZ]) and online outpatient clinic letters were then used to collect data on the patients’ HbA1c and any resulting delay in their elective surgery. Patients with and without DM were included in the study.

The study was registered as a local clinical audit (registration code DIAB-17-18-08) and as such ethical approval was not required. The study was registered with Researchregistry.com, registration number 4756. Reporting of this work followed the STROCSS (Strengthening the Reporting of Cohort Studies in Surgery) criteria. [20].

Results

Demographics and Type of Surgery

Of the 1000 consecutive patients reviewed, the median age was 69 (range 16–92) years, 42% were female and 20% had diabetes (Table 1). There was an even split between vascular and orthopedic surgery patients. The type of surgery performed is summarized in Table 2.

Table 1 Glycemic status and type of diabetes for all 1000 patients in the study
Table 2 Type of surgery performed

HbA1c Concentration

Overall there were 201 patients with DM, among whom 155 (77%) achieved an HbA1c below the target of 69 mmol/mol (8.5%) prior to their surgery. A total of 46 patients (23%) with DM did not meet the recommended preoperative HbA1c target.

Of the 46 individuals with DM who had a preoperative HbA1c ≥ 69 mmol/mol (8.5%), 41 underwent surgery irrespective of their preoperative   glycemic control. The procedures carried out for these individuals are listed in Table 3. The remaining five individuals had their surgery deferred or canceled as a result of suboptimal preoperative glycemic control.

Table 3 Type of surgery performed without deferment on those with glycated hemoglobim concentrations that exceeded 69 mmol/mol

Deferred Surgeries

The details of the five individuals who had their surgery delayed or canceled due to a high HbA1c concentration are summarized in Table 4.

Table 4 Details of patients whose surgery was delayed/canceled due to high glycated hemoglobin concentration

The surgery of one individual who was scheduled to have a vascular procedure for an occluded superior femoral artery stent, for which a femoro-distal bypass was indicated, was canceled due to high HbA1c. However, because the patient had continued poor glycemic control but the symptoms showed an improvement, surgery was no longer deemed urgent, and the decision was made to continue conservative treatment. This individual had the highest HbA1c readings of all the patients included in this study (134 and 140 mmol/mol [14.4 and 15%, respectively]).

Of the seven orthopedic patients with an HbA1c of ≥ 69 mmol/mol, only three ultimately had their respective surgery without delay. The HbA1c values of these three patients were 74, 72, and 69 mmol/mol (8.9, 8.7 and 8.5%), respectively. The patient with the HbA1c of 74 mmol/mol (8.9%) underwent an urgent foot amputation following a diagnosis of osteomyelitis.

No further preoperative assessments or admission details were found for the two patients with high HbA1c concentrations awaiting total hip replacements. These patients may have had their surgery elsewhere or had their surgery canceled altogether, thereby opting for conservative management.

Discussion

Our study shows that of 201 vascular and orthopedic patients with DM, only 46 (23%) had a pre-operative HbA1c ≥ 69 mmol/mol (8.5%). Of these latter 46 patients, only five individuals had their surgery deferred or canceled due to exceeding the target HbA1c concentration, while the majority (41/46) underwent surgery as planned.

Our institution is a 1000-bed tertiary teaching hospital that primarily serves the population of Norfolk, UK as well as neighboring counties. The prevalence of DM (diagnosed and undiagnosed) in Norfolk alone was estimated at 68,098 in 2015 [1]. At the time of the 2017 National Diabetes Inpatient Audit there were 140 inpatients with DM at the NNUH, accounting for 15.9% of the total number of beds included in the audit [21].

It is clear that if poor glycemic control was not picked up by the diabetes caregiver prior to the patient attending the the preoperative assessment clinic, then 2 weeks would be insufficient time to optimize glycemic control and the decision on whether to defer surgery would depend on the outcome of a discussion between the anaesthetist and the surgeon. The predominant factor in this discussion would be the urgency of the surgery. In line with UK national guidelines, the preoperative assessment clinic staff does not manage the diabetes care of individuals other than giving out instructions on how to manage their diabetes medications on the day prior to and on the day of surgery [19]. Among the patients reviewed here, those whose surgery could be safely deferred were asked to return to their primary diabetes caregiver who would then apply the current standard of care to improve glycemic control. These patients were sent to their diabetes caregiver with the information that the delay in surgery was due to poor glycemic control, and they were given additional information on the glycemic target. The vast majority of patients with type 1 diabetes (> 98%) in our study were looked after by secondary care services, while > 80% of those with type 2 diabetes were looked after by primary care services.

We believe that the sample size was large enough to allow for the primary analysis—i.e. to determine how many patients with DM had a pre-operative HbA1c below the target of 69 mmol/mol (8.5%) for elective orthopedic and vascular surgery. We were also able to review any patients who had a delay in their surgery due to high HbA1c. A strength of our study is that it was a single-center study which enabled us to comprehensively collect the data on HbA1c and on whether surgery went ahead or not.

A limitation in our data was that we did not collect information on interventions implemented to optimize glycemic control before the preoperative assessment, a task which may have been performed by primary or secondary care teams [4, 19]. Such data would have demonstrated whether patients with high HbA1c concentrations were being targeted for optimization of glycemic control before being referred for surgery. In addition, the aim of this study was not to look at any excess morbidity and mortality following surgery in patients with an HbA1c ≥ 69 mmol/mol. We are aware that the sample size was too small for such a study, and other work is ongoing to address this question. Further limitations to our study include not collecting data on the use of continuous glucose monitoring or insulin pump use nor on who was at high risk of developing hypoglycemia.

Data are available which show that post-operative outcomes worsen with worsening of preoperative glycemic control [5,6,7,8,9,10,11,12,13,14,15,16,17,18]. This increased postoperative complication rate is also associated with increased costs [22]. However, little data are currently available on whether various preoperative HbA1c targets are achievable for those undergoing elective surgery or whether clinicians adhere to these guidelines before proceeding with surgery. The data from the present study show that the orthopedic department adhered relatively closely to the target HbA1c of 69 mmol/mol. A previous study by Giori et al. focused on whether people with diabetes requiring total joint arthroplasty at a single center in the USA could achieve a pre-operative HbA1c goal of 53 mmol/mol [23]. These authors found that 94% of their patients ultimately did achieve the target HbA1c, but that 15% had their surgery delayed due to poor glycemic control, similar to our findings. These results led them to conclude that a goal designed to reduce perioperative risk should be achievable but also that it may not be possible for some patients with DM to achieve this target HbA1c of < 53 mmol/mol. However, Giori et al. [23] aimed for much tighter glycemic control than the UK current guidelines, which recommend an HbA1c of < 69 mmol/mol [19]. Furthermore, many individuals requiring joint arthroplasty are elderly, and such ‘tight’ glycemic control may be inappropriate and indeed harmful in this patient population [24, 25]. Some authors from the USA have shown that risks of excess postoperative complications first occur when the HbA1c rises to > 58 mmol/mol [26], and thus a target of < 58 mmol/mol may be inappropriate.

The recommended threshold for preoperative HbA1c differs between studies, ranging between 53 and 75 mmol/mol [6, 26,27,28,29,30]. When investigating joint infections following total knee arthroplasty [31] and hip arthroplasty [29], Cancienne et al. calculated that a target HbA1c of 64 or 58 mmol/mol, respectively, could be used. However, for both studies the authors suggested that HbA1c had poor sensitivity as an independent predictor of infection.

There are data to suggest that those people who are hyperglycemic, but not previously known to have diabetes preoperatively, have poorer postoperative outcomes than those known to have diabetes [32,33,34]. These data suggest that for those who are at risk of developing hyperglycemia, pre-operative HbA1c testing may be beneficial [35]. Those people who may benefit from preoperative testing of HbA1c are listed in Table 5.

Table 5 People without diabetes who may benefit from having their pre-operative glycated hemoglobin concentration measured

Conclusions

In summary, our data show that the majority (77%) of patients in our convenience sample of 1000 consecutive patients undergoing elective vascular and orthopedic surgery were able to achieve a target HbA1c of < 69 mmol/mol. The current preoperative guidance is therefore achievable in a real-life setting. However, as is stated in the national guidance, this target should only be used where it is safe to do so and that a degree of clinical discretion is necessary.