Abstract
Introduction
Diabetes is a common co-morbidity of patients undergoing spinal surgery in the UK but there are no published studies from the UK, particularly with respect to length of hospital stay and complications. The aims of this study were to identify complications and length of hospital stay in patients with diabetes undergoing spinal surgery.
Methods
Data were collected retrospectively for 111 consecutive patients with diabetes (and 97 age and sex matched control patients, identified using computer records) who underwent spinal surgery between 2004 and 2010 in a single centre. The data collected included operative time, blood loss, details of surgery, Clavien complications and length of hospital stay.
Results
No significant differences were found by group in operative time, blood loss, instrumentation, use of graft or revision surgery. Overall complication rates were higher in the patients with diabetes than in the controls (28.8% vs 15.5%). The mean hospital stay was significantly longer for patients with diabetes than for control patients (4.6 vs 3.2 days, p<0.001).
Conclusions
This study identified a significantly higher Clavien grade I complication rate and length of hospital stay in patients with diabetes undergoing spinal surgery than control patients (p=0.02). This has resulted in a predictive model being generated. Of note, no infections were seen in patients with diabetes, suggesting that infection rates in this particular group of patients undergoing spinal surgery might not be as high as considered previously.
Keywords: Diabetes, Spinal surgery, Clavien complications, Length of stay
Diabetes is a significant health burden in the UK, affecting 2.9 million people. 1 It is a chronic disease and is associated with both microvascular and macrovascular complications. Impaired leucocyte function, microvascular changes and hyperglycaemia result in an immunocompromised state, 2,3 explaining the increased infection risk seen in patients undergoing spinal surgery. 4,5 It has also been shown that diabetic patients are associated with a longer hospital stay, increased operative time, a higher risk of non-union and overall increased complications. 5–11
Diabetes is a common and significant co-morbidity of patients undergoing spinal surgery. The coexistence of diabetic neuropathy can present a diagnostic challenge and the increased risks associated with surgery in patients with diabetes leads to some reluctance to perform surgery in this group of patients. The aims of this study were to identify the rate of Clavien complications and the length of hospital stay in diabetic patients, compared with control patients, undergoing spinal surgery in a single centre in the UK.
Methods
Computer records were used to identify all patients with diabetes (n=111) who underwent spinal surgery between 2004 and 2010 in a single centre. Control patients were identified using computer records, and were matched for age and sex (n=97). Data were collected retrospectively for these consecutive patients from their medical records, and included demographics, co-morbidity, drug history, type of surgery, operative time, blood loss, postoperative complications, length of hospital stay and need for revision surgery during follow-up.
All patients (both diabetic and non-diabetic patients) received antibiotics at induction of anaesthesia, according to local microbiology guidelines. Surgery was carried out in a standard fashion with midline and lateral decompression through a midline approach. Either a laminotomy or a laminectomy was carried out depending on the pattern of stenosis and foraminal decompression was carried out as necessary. Fusion (usually with instrumentation) was carried out where there was evidence of or risk of developing instability such as spondylolisthesis of the spine. In multiple level surgery, only unstable levels were fused.
Complications were classified according to Clavien grade, 12,13 which is detailed in Table 1. Statistical analysis was performed using chi-squared tests for categorical variables and Mann–Whitney U tests for continuous variables. Regression models were used to predict the expected length of hospital stay for each group (diabetics vs controls) as a function of the patient’s age. The model selection procedure was based on the Bayesian information criterion. 14 All calculations were performed in using R statistical software version 2.12.2 (R Foundation for Statistical Computing, Vienna, Austria).
Table 1.
Details of Clavien classification of complications
| Grade | Definition |
|---|---|
| I | Any deviation from the normal postoperative course without the need for pharmacological treatment or surgical, endoscopic and radiological interventions. Allowed regimens are: antiemetics, antipyretics, analgesics, diuretics, electrolytes and physiotherapy. Includes wound opened at the bedside. |
| II | Requiring pharmacological treatment with drugs other than those allowed for grade I. Includes blood transfusions and total parental nutrition. |
| III | Requiring surgical, endoscopic or radiological intervention |
| IIIa | Intervention not under general anaesthesia |
| IIIb | Intervention under general anaesthesia |
| IV | Life threatening complication requiring intensive care unit management |
| IVa | Single organ dysfunction (including dialysis) |
| IVb | Multiorgan dysfunction |
| V | Death |
Results
We found no significant differences by group in age, sex, revision procedure, instrumentation, use of graft (all autografts, predominantly from the iliac crest), operative time or blood loss (Table 2). Of note, the most frequently performed procedure in both groups was lumbar decompression (72.1% and 76.5% in controls and patients with diabetes respectively), followed by lumbar instrumented fusion (12.4% and 10.9% respectively). Furthermore, the majority of patients had single level surgery (75.3% and 72.1% in controls and patients with diabetes respectively). Of the 111 patients with diabetes, 16 had type 1 diabetes (14.4%). The mean length of hospital stay was significantly longer for patients with diabetes than for control patients (4.6 vs 3.2 days, p<0.001) (Fig 1). For both controls and patients with diabetes, the length of stay also correlated positively with age (p<0.001).
Table 2.
Pre and perioperative characteristics for control and diabetic patients (with ranges or percentages)
| Controls (n=97) | Diabetics (n=111) | p-value | |
|---|---|---|---|
| Sex | 0.7928 | ||
| Male | 56 (57.7%) | 61 (55.0%) | |
| Female | 41 (42.3%) | 50 (45.0%) | |
| Age at surgery (years) | 63 (31–86) | 62 (31–86) | 0.9613 |
| Revision procedure | 16 (16.5%) | 24 (21.6%) | 0.4475 |
| Types of surgery | |||
| Lumbar discectomy | 9 (9.2%) | 6 (5.4%) | |
| Lumbar decompression | 70 (72.1%) | 85 (76.5%) | |
| Lumbar fusion | 1 (1.1%) | 3 (2.7%) | |
| Lumbar instrumented fusion | 12 (12.4%) | 12 (10.9%) | |
| Lumbar inspacer | 0 (0%) | 1 (0.9%) | |
| Coccygectomy | 0 (0%) | 1 (0.9%) | |
| Anterior cervical discectomy and fusion | 4 (4.1%) | 2 (1.8%) | |
| Thoracic decompression | 1 (1.1%) | 1 (0.9%) | |
| Instrumentation | 15 (15.5%) | 17 (15.3%) | 0.4167 |
| Number of levels | 0.3241 | ||
| 1 | 73 (75.3%) | 80 (72.1%) | |
| 2 | 17 (17.5%) | 25 (22.5%) | |
| 3 | 6 (6.2%) | 5 (4.5%) | |
| 4 | 0 (0%) | 1 (0.9%) | |
| 5 | 1 (1.0%) | 0 (0%) | |
| Use of graft | 12 (12.4%) | 15 (13.5) | 0.9699 |
| Operative time (mins) | 100 (45–355) | 105 (24–323) | 0.5324 |
| Blood loss in (ml) | 158 (0–3,500) | 148 (0–2,265) | 0.7400 |
| Hospital stay (days) | 3.2 (1–28) | 4.6 (1–22) | <0.001 |
Figure 1.
Length of hospital stay for control patients and patients with diabetes
Of the multiple regression models used, the Sichel model yielded the best fit. The expected values of length of stay predicted by the model differed by group (p=0.002) at 2.5 days for controls and 3.5 days for patients with diabetes. Each year increment in the patient’s age at surgery increased the length of stay significantly by 1.5% (p=0.017), regardless of whether they were a control patient or a patient with diabetes. Figure 2 shows the expected number of days in hospital as predicted by the Sichel model as a function of the patient’s age by group.
Figure 2.
Sichel model showing expected length of hospital stay following spinal surgery by age for control patients and patients with diabetes, including 95% prediction confidence intervals
The rate of complications was significantly higher in patients with diabetes than in controls (32/111 [28.8%] vs 15/97 [15.5%], p=0.01848) (Table 3). However, rates of complication requiring readmission were higher in controls than in patients with diabetes (9.3% vs 5.4%). Patients with diabetes had significantly more Clavien grade I complications than the controls (p=0.02), with higher rates of urinary retention (n=19 vs n=8). Interestingly, no infections were identified in the diabetes group although four infections were identified in the control group. (Two of these patients had to return to theatre for washout.)
Table 3.
Details of Clavien complications for diabetic and control patients
| Controls (n=97) | Diabetics (n=111) | |
|---|---|---|
| No complications | 81 (83.5%) | 79 (71.2%) |
| Clavien I | 9 | 25 |
| Urinary retention | 8 | 19 |
| Ileus | 1 | 1 |
| Basal atelectasis | 1 | |
| Wound dehiscence (managed conservatively) | 1 | |
| Prolonged wound discharge | 1 | |
| Neurapraxia | 1 | |
| Constipation | 1 | |
| Clavien II | 4 | 4 |
| Atrial fibrillation (requiring bisoprolol) | 1 | |
| Pneumonia | 1 | 1 |
| Urinary tract infection | 1 | 1 |
| Blood transfusion | 1 | |
| Wound infection | 2 | |
| Clavien IIIa | 0 | 1 |
| Myocardial infarction, coronary artery stent | 1 | |
| Clavien IIIb | 2 | 2 |
| Washout of deep wound infection | 2 | |
| Evacuation of haematoma | 1 | |
| Reoperation due to malplacement of pedicle screw | 1 | |
| Total complications | 15 (15.5%) | 32 (28.8%) |
The level of glycaemic control (HbA1C) and medications used to treat diabetes in these patients are shown in Table 4. Overall, the average HbA1C was 7.1% (range: 4.6–10.0%) and HbA1C increased with level of treatment for diabetes. Although no correlation between HbA1C and complications was identified, there was a trend towards an increased length of stay with increased HbA1C (Pearson correlation coefficient r=0.74). Of the 16 patients with type 1 diabetes, 3 patients had complications (2 had Clavien grade I and 1 had Clavien grade II).
Table 4.
Details of diabetic treatments and corresponding HbA1 c values for the diabetic patients
| Treatment | Number of patients (n=111) | HbA1c (range) |
|---|---|---|
| Diet controlled | 18 (16.2%) | 6.4% (5.9–7.6%) |
| Oral hypoglycaemic/s | 59 (53.2%) | 6.8% (4.6–8.3%) |
| Insulin | 20 (18.0%) | 7.6% (5.5–10.0%) |
| Insulin + oral hypoglycaemic | 14 (12.6%) | 8.3% (6.9–10.0%) |
Discussion
To our knowledge, this is the first study to look at the rates of complications and length of hospital stay in diabetic patients undergoing spinal surgery in the UK. Furthermore, these consecutive patients all underwent surgery in a single centre, and age and sex matched controls were identified to allow comparisons to be made. Limitations include the retrospective nature of the study, the lack of functional outcomes data and the relatively low numbers of patients in each group.
Our overall complication rates for both diabetic and control patients (28.8% and 15.5% respectively) are in line with other published series (Table 5). 4–7,9,11,15–17 Reports from the existing literature show increased rates of complications, particularly infection, in patients with diabetes following spinal surgery, leading to some reluctance by spinal surgeons to offer surgery to these patients. In our centre, the threshold for offering spinal surgery is identical in patients with or without diabetes. However, we do advise patients with diabetes preoperatively that they are likely to be at a higher risk of complications overall based on the previously published literature.
Table 5.
Published studies of complications following spinal surgery in patients with diabetes
| Study | Patients with diabetes | Controls | Infection rate for diabetics | Infection rate for controls | Overall complication rate for diabetics | Overall complication rate for controls |
|---|---|---|---|---|---|---|
| Pull ter Gunne 3 | 290 | 2,884 | 7.6% | 3.8% | Not given | Not given |
| Glassman 4 | 94 | 43 | 10.6% | 4.7% | 54.2% | 20.9% |
| Browne 5 | 11,135 | 186,326 | 0.7% | 0.3% | 19.0% | 13.4% |
| Cook 6 | 3,432 | 34,300 | 0.3% | 0.3% | 10.0% | 15.2% |
| Chen 8 | 30 | 165 | 30.0% | 11.0% | Not given | Not given |
| Freedman 10 | 199 | 2,207 | 10.0% | 4.0% | 8.0% | 6.0% |
| Simpson 14 | 62 | 62 | 34.0% | 0.0% | Not given | Not given |
| Kawaguchi 15 | 18 | 34 | 16.7% | 0.0% | Not given | Not given |
| Bendo 16 | 32 | 0 | 6.3% | – | 31.0% | – |
Although a higher overall complication rate was identified in patients with diabetes than in controls, the majority of these complications were only Clavien grade I (and most frequently urinary retention), with the absence of any infections highlighting the fact that surgery could be offered more readily to these patients than is currently the case in some centres. Indeed, several other large studies have identified low rates of infections in patients with diabetes, confirming our findings.
Nevertheless, the discrepancy in infection rates between published studies could also be due to differences in the rates of instrumentation. For example, in a study of 137 patients undergoing posterior instrumented fusion, with 94 patients with diabetes, the infection rate in the patients with diabetes was 10.64% and the overall complication rate in this subgroup was 54.6%. 5 However, in another study of 32 patients with diabetes undergoing posterior instrumented fusion, the infection rate was 6.3% and the overall complication rate was 31.0%. 17 In our study, the rates of instrumentation were low (15.3% in diabetics and 15.5% in controls), which could be a possible explanation for the absence of any infections in the diabetes group. A much larger study would be needed to address the relationship between type of spinal surgery, diabetes and risk of complications.
Another explanation for discrepancies in published infection rates could be differences in diabetic control (ie HbA1C levels). In our study, HbA1C was higher for patients taking insulin and an oral hypoglycaemic than for those who were diet controlled (Table 4), as expected. The average HbA1C of our patients was 7.1% and it has been shown that patients with HbA1C levels below 7% have lower rates of postoperative infections. 18 It has also been demonstrated that there is a negative correlation between the rate of recovery after spinal surgery and preoperative HbA1c, 16 and that patients with higher preoperative HbA1C have longer hospital stays (associated with higher costs) following spinal surgery. 19
In the present study, there was a correlation between HbA1C and length of stay but this was not statistically significant and a larger study would be needed to address this. Furthermore, the number of patients with type 1 diabetes in this study was too small to identify any significant relationship with risk of complications versus patients with type 2 diabetes. Again, larger studies would be needed to look at this in more detail.
Interestingly, significantly higher rates of postoperative urinary retention in patients with diabetes were observed compared with controls (8.2% vs 17.1%), which was not attributable to urinary tract infection, suggesting a neuropathic origin. It is known that patients with diabetes can have autonomic and peripheral neuropathy, affecting the sensory nerve supply of the bladder, and this is unmasked after general anaesthesia, leading to urinary retention. Similar rates of urinary retention have been observed previously (4.7% and 13.8% in patients with diabetes and controls respectively), 5 suggesting it could be beneficial to catheterise these patients intraoperatively.
A significantly longer length of hospital stay was observed in the patients with diabetes (4.6 days vs 3.2 days). This could be explained by the increased overall complication rate observed in the diabetic group although diabetes has been identified previously as an independent risk factor for longer hospital stay. 19 Furthermore, patients with diabetes were required to demonstrate adequate glycaemic control postoperatively prior to discharge, which could have further increased their length of stay. Additionally, using the Sichel regression model, we developed a predictive tool for length of hospital stay for patients with and without diabetes, based on age, which can be used both in preoperative discussions with patients and for purposes of financial planning.
Conclusions
This single-centre UK study has identified a significantly higher Clavien grade I complication rate and length of hospital stay in patients with diabetes undergoing spinal surgery compared with control patients. Of note, no infections were seen in the diabetes group, suggesting that infection rates in patients with diabetes undergoing spinal surgery might not be as high as considered previously. In addition, we have generated a predictive model based on these data. This can be used to inform patients preoperatively about length of stay and also for financial planning by hospital departments. Further studies will be required to validate this model in clinical practice.
References
- 1.Diabetes UK. Diabetes in the UK. London: Diabetes UK; 2012 [Google Scholar]
- 2.Delamaire M, Maugendre D, Moreno Met al Impaired leucocyte functions in diabetic patients. Diabet Med 1997; 14: 29–34 [DOI] [PubMed] [Google Scholar]
- 3.Silhi N. Diabetes and wound healing. J Wound Care 1998; 7: 47–51 [DOI] [PubMed] [Google Scholar]
- 4.Pull ter Gunne AF, Cohen DB. Incidence, prevalence, and analysis of risk factors for surgical site infection following adult spinal surgery. Spine 2009; 34: 1,422–1,428 [DOI] [PubMed] [Google Scholar]
- 5.Glassman SD, Alegre G, Carreon Let al Perioperative complications of lumbar instrumentation and fusion in patients with diabetes mellitus. Spine J 2003; 3: 496–501 [DOI] [PubMed] [Google Scholar]
- 6.Browne JA, Cook C, Pietrobon Ret al Diabetes and early postoperative outcomes following lumbar fusion. Spine 2007; 32: 2,214–2,219 [DOI] [PubMed] [Google Scholar]
- 7.Cook C, Tackett S, Shah Aet al Diabetes and perioperative outcomes following cervical fusion in patients with myelopathy. Spine 2008; 33: E254–E260 [DOI] [PubMed] [Google Scholar]
- 8.Veeravagu A, Patil CG, Lad SP, Boakye M. Risk factors for postoperative spinal wound infections after spinal decompression and fusion surgeries. Spine 2009; 34: 1,869–1,872 [DOI] [PubMed] [Google Scholar]
- 9.Chen S, Anderson MV, Cheng WK, Wongworawat MD. Diabetes associated with increased surgical site infections in spinal arthrodesis. Clin Orthop Relat Res 2009; 467: 1,670–1,673 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Schimmel JJ, Horsting PP, de Kleuver Met al Risk factors for deep surgical site infections after spinal fusion. Eur Spine J 2010; 19: 1,711–1,719 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Freedman MK, Hilibrand AS, Blood EAet al The impact of diabetes on the outcomes of surgical and nonsurgical treatment of patients in the spine patient outcomes research trial. Spine 2011; 36: 290–307 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Clavien PA, Sanabria JR, Strasberg SM. Proposed classification of complications of surgery with examples of utility in cholecystectomy. Surgery 1992; 111: 518–526 [PubMed] [Google Scholar]
- 13.Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004; 240: 205–213 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kuha J. AIC and BIC: comparisons of assumptions and performance. Sociol Methods Res 2004; 33: 188–229 [Google Scholar]
- 15.Simpson JM, Silveri CP, Balderston RAet al The results of operations on the lumbar spine in patients who have diabetes mellitus. J Bone Joint Surg Am 1993; 75: 1,823–1,829 [DOI] [PubMed] [Google Scholar]
- 16.Kawaguchi Y, Matsui H, Ishihara Het al Surgical outcome of cervical expansive laminoplasty in patients with diabetes mellitus. Spine 2000; 25: 551–555 [DOI] [PubMed] [Google Scholar]
- 17.Bendo JA, Spivak J, Moskovich R, Neuwirth M. Instrumented posterior arthrodesis of the lumbar spine in patients with diabetes mellitus. Am J Orthop 2000; 29: 617–620 [PubMed] [Google Scholar]
- 18.Dronge AS, Perkal MF, Kancir Set al Long-term glycemic control and postoperative infectious complications. Arch Surg 2006; 141: 375–380 [DOI] [PubMed] [Google Scholar]
- 19.Walid MS, Zaytseva N. How does chronic endocrine disease affect cost in spine surgery? World Neurosurg 2010; 73: 578–581 [DOI] [PubMed] [Google Scholar]