Complications after hip fracture surgery: are they preventable?

A prospective cohort study was conducted at the Departments of Trauma Surgery and Orthopedic Surgery of University Medical Center Groningen (UMCG) in the Netherlands. Informed consent was obtained from all patients and the study was approved by the Medical Ethical Committee of UMCG (METc 2011/164). The study was conducted in accordance with the Declaration of Helsinki.

All patients aged 60 years or older with a hip fracture treated at UMCG between 1 July 2009 and 1 May 2013 were included in this study. Patients were excluded if they had multiple injuries caused by high energetic trauma. A hip fracture was defined as a femoral neck fracture (dislocated or not dislocated) or a trochanteric fracture [subdivided into type A1, A2 or A3 according to the Arbeitsgemeinschaft für Osteosynthesefragen (AO) comprehensive classification]. In 2009, a comprehensive care pathway for the treatment of hip fractures was implemented at UMCG. The care pathway has been described before [10]. This comprehensive multidisciplinary care pathway was developed to include all elements of care in the trajectory, from arrival in the emergency room to the moment of discharge from the rehabilitation unit of the nursing home. This operative protocol for hip fractures at UMCG is based on the Dutch surgical guideline for hip fractures [11].

The electronic files of the included patients were searched for complications occurring from admittance up to the last outpatient visit 6 months after surgery. Registered baseline characteristics were demographic information, medical history, fracture classification, trauma mechanism, living situation, American Society of Anesthesiologists (ASA) classification and baseline blood values (hemoglobin and electrolytes). The Charlson comorbidity index (CCI) was calculated. This score is a method for estimating the risk of death based on comorbidities [12]. To calculate the score, 22 comorbid conditions are assigned a score of 1, 2, 3 or 6, depending on the risk of dying associated with each one. The index is the sum of the scores [12]. During hospitalization type of implant, waiting time to surgery, operation time, type of anesthesia, preoperative fasting time, length of hospital stay and blood values at discharge were taken from the electronic hospital registration system. Clinical complications and side effects were registered. All complications were also registered up to 6 months postoperatively; the mortality rate was followed up to 1 year after the operation.

Complications were divided into minor and major complications. Examples of minor complications are: hematoma, urine retention, electrolyte disturbances and leg length discrepancy. Examples of major complications are: all reoperations, delirium, pneumonia, cardiac and neurological complications, and pressure ulcers.

Descriptive statistics were used to describe the characteristics of the study population at baseline and the number and type of complications after surgery. Logistic regression analysis was used to investigate which variables were significant risk factors for the occurrence of perioperative complications. In this analysis, the dependent variable was complication (yes/no) and the independent variables were age, gender, fracture type, CCI, pre-fracture living situation, >1 day delay to first procedure (yes/no) and type of anesthesia (spinal vs. general). For the analysis patients were divided into four age groups (60–65, 65–75, 75–85, >85), four groups of CCI (0, 1, 2, ≥3), and four groups of living situation (living independently, independently with others, in an assisted living facility, in a nursing home).

Additional analyses were conducted to investigate whether the independent variables were significant risk factors for several specific complications (delirium, congestive heart failure and pneumonia). In the logistic regression analyses, the independent variables were first introduced separately in univariate models, after which multivariate models were constructed with the variables with p < 0.20 using a stepwise forward selection method. Variables with p ≤ 0.10 were retained in the final multivariate models. Odds ratios (OR) with the corresponding 95% confidence intervals (CI) were calculated for each independent variable.

Cox proportional hazards regression analyses were performed to estimate the association between several independent variables (age, gender, fracture type, complication, etc.) and 1-year mortality. First, univariate regression models were created, after which the variables with p < 0.20 were eligible to be included in the multivariate regression models. A stepwise forward selection method was used. Variables with p ≤ 0.10 were retained in the final multivariate models. Hazard ratios (HR) with the corresponding 95% CI were calculated. All statistical analyses were performed using SPSS Statistics for Windows (Version 22.0, Armonk, NY: IBM Corp.). p < 0.05 was considered to indicate statistical significance.

A total of 359 patients (75%) suffered one or more complications during the 6 months of follow-up; 210 patients (44%) suffered multiple complications (Table 2). Only 119 patients (25%) did not have any negative side effects from the treatment. Of all complications, 173 (48%) were minor and 186 (52%) were classified as major. The three most frequently encountered medical complications were delirium (N = 98, 20%), pneumonia (N = 47, 10%) and congestive heart failure (N = 25, 5%). Pressure ulcers were registered in five patients (1%). Complications related to the operative procedure were seen in 13% of the patients (total wound and implant problems N = 63). Reoperation was performed in 52 patients (11%) during the follow-up period. Eighty patients (16%) only had complications that were registered during outpatient visits; most seen complications at this time were leg length difference and persistent pain. After 1 year, 129 patients (27%) had died.

A statistically significant association was found between a CCI score ≥3 and the occurrence of complications after hip fracture surgery [OR = 3.19, 95% CI 1.67–6.08, p < 0.001 (Table 3)]. There was no statistically significant association between other independent variables and the occurrence of complications. Patients who received general instead of spinal anesthesia did have more complications, but this difference did not reach statistical significance (OR = 1.51, 95% CI 0.97–2.35, p = 0.07).

Delirium was the most frequently observed complication. Not living independently, higher age and delay of surgery of more than 1 day were all statistically significant risk factors for getting delirium (Table 4). There was only one independent risk factor for congestive heart failure: CCI ≥ 3 (OR = 16.49, 95% CI 2.13–127.57, p = 0.007). For postoperative pneumonia, the relationship with CCI was borderline significant (OR = 2.18, 95% CI 0.97–4.90, p = 0.06). There was no statistically significant association between postoperative pneumonia and the other variables.

Patients with a complication did not have a higher mortality rate (OR = 1.18, 95% CI 0.70–2; p = 0.52, Table 5). In univariate analyses, only CCI and higher age were found to be statistically significant predictors of mortality. However, when independent variables, including the various complications, were investigated in multivariate models, delirium as well as pneumonia increased mortality risk after hip fracture surgery [HR = 2.20, 95% CI 1.46–3.32, p < 0.001; and HR = 2.05, 95% CI 1.24–3.39, p = 0.005, respectively (Table 6)].

The three most encountered medical complications were delirium, pneumonia and congestive heart failure. A delirium incidence of 20% was found in this study. The reported incidence of delirium varies widely between 4 and 53% [15]. The comprehensive care pathway followed in this study included consultation with a geriatrician. There is some evidence that delirium can be prevented or that its severity can be reduced after geriatric consultation [16]. On the other hand, a geriatrician will diagnose delirium sooner than other specialists, resulting in a higher yet more reliable incidence rate.

Pneumonia and congestive heart failure were the second and third most prevalent medical complications, with an incidence of, respectively, 10 and 5%. This is in accordance with other research [3, 17]. Possible preventive measures for these complications that were already included in the protocol for the care pathway is this study were investigation of cardiac status by the anesthesiologist, surgery on the first day after admission in supine position and early mobilization.

All surgical complications together formed 13% of the complications. Because of the heterogeneity of these complications and the small numbers of each of its different components (hematoma, infection, different implant problems), this was not further analyzed. The incidence of surgical complications matches the literature. The reported surgical site infection rate after hip fracture surgery is about 5% for orthopedic trauma surgery [18, 19]. The incidence of reoperation depends on fracture type and implant, but is reported between 7 and 38% in femoral neck fractures [20, 21].

The percentage of pressure ulcers in our study was low (1%), and much lower than reported in the literature; a recent study conducted in the US showed that about one-third of hip fracture patients developed a pressure ulcer during follow-up [22]. An explanation for the low incidence could be the special preventive measures that were taken in this study. Already on the ER patients with a hip fracture were placed on a pressure ulcer-preventive mattress, and almost all patients were operated on within one day. However, the number of pressure ulcers might have been underestimated as admission time to the hospital was short, and probably not every ulcer was reported by the nursing home physician or seen during physical examination at the outpatient clinic.

In this study, the only significant risk factor of developing a complication was having a high CCI (≥3). Other studies also show that having more comorbidities is the strongest risk factor for complications [3, 6]. Risk factors close to statistical significance are having an A3 type intertrochanteric fracture and getting general instead of spinal anesthesia. The A3 type intertrochanteric fracture is classified as unstable and more difficult to treat than other trochanteric fractures. Patients seem to need more analgesics [23], which might lead to more complications. Type of anesthesia is a modifiable risk factor. In this study, the number of patients that underwent general anesthesia was high (65%). Whether general or spinal anesthesia leads to fewer complications in hip fracture patients is still a matter of debate [24‐27]. The recent Cochrane review did not find a difference in either technique and asked for more research [28]. Most important factor seems to be the patient himself: for some cardiac conditions general anesthesia is more safer, for other pulmonary or neurological conditions spinal anesthesia is preferred.

Another known, possibly preventable risk factor for complications is delay of surgery for more than 2 days [6]. This effect was not seen in the present study, perhaps because one part of the care pathway was to operate on all patients the morning after admission, so only for a small number of patients (9%) was the operation postponed longer than 2 days. Surgery was only postponed for strict medical reasons such as sepsis or fulminant pneumonia; in these cases, these medical reasons were considered comorbidities, not complications, as they were already present at arrival in the emergency room.

The risk for delirium was higher in older patients and those living in a nursing home; some of these patients probably had cognitive dysfunction already (in this study population this was not measured during admittance), and these patients are known to be at risk for delirium [29]. In general, patients living in a nursing home have a lower performance status in comparison to patients living independently; mostly their CCI is higher resulting in a higher complication rate, as shown especially in delirium. Social isolation might play a role as a risk factor for delirium. However, a definitive answer cannot be given. Patients living independently may live as much or even more in social isolation as patients in nursing homes do. In the literature there is some evidence that the presence of family members in the hospital will help to reduce the incidence of delirium [30], moreover, that social isolation leads to a prolonged hospitalization [31].

The risk factors for delirium found in this study are also in line with the British NICE guidelines for delirium, with the most important risk factors attributed to older age, cognitive impairment and having a hip fracture [32]. The only preventable risk factor for delirium known in the literature is delay in surgery for more than 1 day, resulting in a prolonged period of bed rest and pain. We therefore believe it should be a goal of care pathways to operate patients with a hip fracture within a day of admission. In accordance with the recent literature [33], in our study general anesthesia was not associated with a higher risk of delirium.

Early mobilization is one of the aspects of the comprehensive care pathway. Research has shown that early mobilization is probably one of the most important measures to reduce the incidence of pneumonia [34]. This study did not show other preventable independent risk factors for pneumonia or congestive heart failure.

The 1-year mortality rate was 27%. Although most studies report mortality rate after 3 or 4 months, this annual rate is in accordance with the literature [3, 8]. Mortality was not influenced by complications in general. The only independent risk factors for mortality were having delirium or pneumonia postoperatively. It is generally assumed that having delirium is a poor prognostic sign, and Witlox et al. showed in their meta-analysis that delirium has been proven to be an independent risk factor not only for mortality but also for institutionalization and dementia [35]. This stresses the importance of reducing the risk of delirium. Higher postoperative mortality after pneumonia has been described before [3] and is probably partly due to poor physical condition, which hampers mobilization.