Inglês (pdf)
Artigo em XML
Referências do artigo
Tradução automática
Enviar este artigo por email
Citado por SciELO 
Similares em
SciELO 
Permalink
INTRODUCTION
Physiological changes inherent to aging predispose to fractures and their complications, which are related to unfavorable outcomes like death1. Worldwide, elderly mortality is high one year after hip fractures, with rates ranging from 22.8% to 29.5%2,3. In Brazil, accessing public hospitals poses challenges, and there are also delays in surgery for these cases4. These delays may heighten the risk of death during hospitalization and within one year, particularly in the Brazilian Public Health System (Sistema Único de Saúde - SUS) hospitals5,6. This can be particularly relevant in areas where transportation is difficult, such as the Amazon Region, where many roads are unpaved, ambulances are insufficient, and riverside populations remain isolated due to the lack of quick river transportation for these situations.
Despite the widely available evidence about mortality after hip fracture worldwide, the studies are usually based on areas of high urbanization2,5,6. In contrast, no data on death rates is available specifically in the Amazon Region. Therefore, formulating evidence-based public health policies becomes a challenge.
Nevertheless, hip fractures are frequent in Amazon. According to the national public health system register, only 1,825 hip fractures were notified between January 2015 and December 2016 in Pará State7. Given the gap in the literature, this study aimed to investigate the factors associated with death up to one year after hip fracture surgery among the elderly admitted to an orthopedic referral hospital in Belém, the capital of Pará, northern Brazil. We hypothesize that associated comorbidities and the time interval between fracture and surgery occurrence influence the patients' survival in one year.
MATERIALS AND METHODS
STUDY DESIGN AND SETTING
This is a retrospective cohort study based on data from medical records. Data were collected from all consecutive patients admitted with hip fracture in Maradei Hospital, in the Belém City, Pará State, Brazil. The hospital is a metropolitan and regional referral center for orthopedic and trauma treatment in Pará. It serves SUS and private individuals through insurance policies, functioning as a teaching hospital for medical students from Federal University of Pará (UFPA).
All patients from SUS were referred to the hospital through Pará's health regulatory system. Those from private insurance were admitted directly from the hospital's emergency department.
ETHICS
The Research Ethics Committee of the Institute of Health Sciences at UFPA approved the study on November 27, 2018, under protocol CAAE 91758418.7.0000.0018. Informed consent was waived, as the study was based mainly on data from medical records; besides, anonymity was guaranteed to the patient and family in cases where a telephone call was accomplished.
PARTICIPANTS AND STUDY SIZE
The hospital admitted patients from the capital, Belém, and its metropolitan region, as well as from the northeast and southeast of the state and from Marajó Island (Figure 1), which were referred by the public health system. Across Pará State, patients from the Tapajós region were the only ones not referred to the hospital. All consecutive elderly patients who received surgical treatment for fractures in the proximal third of the femur (neck, intertrochanteric, and subtrochanteric), between January 1, 2015, and December 31, 2016, were included in the study. The initial date is the inception of the electronic medical records system in the hospital. Therefore, this study used a convenience sample of all consecutive patients for whom it was possible to retrieve data. All patients aged 60 years old or older were considered as elderly. Medical records without an exact surgery date and those treated conservatively were excluded from the study.
VARIABLES AND DATA SOURCES
Possible death in the first year after surgery was verified within a death certificates database provided by the Health Department of the State of Pará (Secretaria do Estado de Saúde do Pará - SESPA) for this study. The dates and causes of death were registered. The cause of death was recorded using an International Statistical Classification of Diseases and Related Health Problems Code (ICD10).
Additional information was also gathered from the medical records, including demographics (age at surgery, sex, and origin), general clinical characteristics (comorbidities, smoking status, date of admission, date of hospital discharge, and date of death if occurred), as well as clinical details related to the hip fracture (affected side of fracture and fracture site, type of surgery, date of fracture, and date of surgery).
The following comorbidities were registered: systemic arterial hypertension, diabetes mellitus, history of stroke, other heart diseases, Alzheimer's disease, and prostate cancer. The presence of pulmonary comorbidities was also verified, such as pulmonary fibrosis, chronic obstructive pulmonary disease (COPD), asthma, and hyper and hypothyroidism. Patients who used at least one smoked tobacco product at least once a week were considered current tobacco users. Patients who declared they did not smoke for at least one year were identified as former smokers.
The proportion of patients undergoing surgery within 48 h of the fracture was evaluated by analyzing the trauma and the surgery dates. This timeframe is considered optimal for minimizing postoperative risks of mortality8,9. Clinical and demographic variables associated with surgery delay (operations taking place after the window of 48 h from the trauma) were investigated.
SOURCES OF BIAS
To mitigate the risk of clerical errors in death certificates, we attempted to confirm the cause and date of death by making telephone contact with the families of the deceased. All data collected for this study were recorded by one researcher and independently double-checked by another.
STATISTICAL ANALYSIS
The binary variables in all patients were described using absolute and relative frequencies. Quantitative characteristics used summary measures (mean, standard deviation, median, minimum, and maximum)10. According to each binary variable, mortality within one year was characterized in absolute and relative frequencies, and the association was verified using chi-square tests or exact tests (Fisher test or likelihood ratio test)10. Quantitative variables were described according to one-year mortality and were compared using the Student's t-test or Mann-Whitney test10. In addition, the unadjusted odds ratios (OR) of each variable of interest with one-year mortality, along with the respective 95% confidence intervals, were estimated in bivariate logistic regression11.
A multiple logistic regression model was performed11, with variables that presented a descriptive level of less than 0.20 (P < 0.20) in the bivariate analyses entering the multiple model, keeping all the variables selected in the model, the "full model", to jointly assess the characteristics that influenced one-year mortality. To avoid multicollinearity, the variable "days between fracture and surgery" was selected instead of the variables "days between fracture and hospitalization" and "days between hospitalization and surgery" since the chosen variable encompasses the other two times.
Microsoft Excel 2003 and IBM-SPSS for Windows v20.0 software were used to tabulate and analyze the data. A significance level of 5% was considered.
RESULTS
Over the two-year study period, the hospital admitted 814 elderly patients for hip fracture treatment, encompassing femoral neck, intertrochanteric, and subtrochanteric regions of the hip. Forty-eight patients for whom the fracture date was unavailable were excluded, along with 224 who were treated conservatively. The remaining 542 cases were evaluated for this study.
Most patients were female (64.9%) and 80 years old or older (55.7%), with the mean age as 80 years (standard deviation - SD of 9.5 years; minimum of 60, maximum of 112 years). When admitted to the hospital, 236 patients (43.5%) were under treatment for hypertension, 110 (20.3%) for diabetes, and 30 (5.5%) suffered from other cardiovascular diseases. At the moment of fracture, 81 (14.9%) were smokers (Table 1).
Table 1 - Epidemiological characteristics of all evaluated patients who treated hip fracture at Maradei Hospital, Pará State, Brazil, 2015-2016
| Variables | Patients | |
|---|---|---|
| N | % | |
| Age (years old) | ||
| 60 to 69 | 88 | 16.2 |
| 70 to 79 | 152 | 28.0 |
| 80 to 89 | 227 | 41.9 |
| ≥ 90 | 75 | 13.8 |
| Sex | ||
| Female | 352 | 64.9 |
| Male | 190 | 35.1 |
| Laterality | ||
| Right | 253 | 46.7 |
| Left | 289 | 53.3 |
| Controlled systemic arterial hypertension | ||
| No | 306 | 56.5 |
| Yes | 236 | 43.5 |
| Controlled diabetes mellitus | ||
| No | 432 | 79.7 |
| Yes | 110 | 20.3 |
| Other cardiovascular disease | ||
| No | 512 | 94.5 |
| Yes | 30 | 5.5 |
| Smoker | ||
| No | 461 | 85.1 |
| Yes | 81 | 14.9 |
| Stroke history | ||
| No | 520 | 95.9 |
| Yes | 22 | 4.1 |
| Alzheimer's disease | ||
| No | 515 | 95.0 |
| Yes | 27 | 5.0 |
| Pulmonary comorbidity | ||
| No | 527 | 97.2 |
| Yes | 15 | 2.8 |
| Prostate cancer history | ||
| No | 539 | 99.4 |
| Yes | 3 | 0.6 |
| Current hyperthyroidism | ||
| No | 541 | 99.8 |
| Yes | 1 | 0.2 |
| Current hypothyroidism | ||
| No | 539 | 99.4 |
| Yes | 3 | 0.6 |
The most common type of fracture was intertrochanteric (63.1%). Most patients were treated with cephalomedullary nails (48.0%) or bipolar hip hemiarthroplasty (27.3%). As shown in table 2, most patients lived outside the capital (59.0%), where the hospital is located (59%), and the mean time between the fracture and hospital admission was 8.7 days (SD 10.7), with 12.4 mean days between fracture and surgery (SD 12.3). The time between fracture and surgery was below 48 h for 20 (3.7%) patients.
Table 2 - Orthopedic procedure to treat hip fracture of all evaluated patients at Maradei Hospital, Pará State, Brazil, 2015-2016
| Variables | Patients | |
|---|---|---|
| N | % | |
| Fracture site | ||
| Femoral neck | 172 | 31.7 |
| Subtrochanteric | 28 | 5.2 |
| Intertrochanteric | 342 | 63.1 |
| Type of surgery | ||
| Cephalomedullary nail | 260 | 48.0 |
| Bipolar hip hemiarthroplasty | 148 | 27.3 |
| Total hip arthroplasty | 19 | 3.5 |
| Dynamic hip screw plate | 90 | 16.6 |
| External fixator | 2 | 0.4 |
| Cannulated screws | 4 | 0.7 |
| Dynamic condylar screw plate | 18 | 3.3 |
| Girdlestone | 1 | 0.2 |
| City | ||
| Belém (capital) | 222 | 41.0 |
| Countryside | 320 | 59.0 |
| Death in one year | ||
| No | 476 | 87.8 |
| Yes | 66 | 12.2 |
| Days between fracture and hospital admission | ||
| Mean and standard deviation | 8.7 ± 10.7 | |
| Median (minimum; maximum) | 6 (- ; 108) | |
| Days in hospital | ||
| Mean and standard deviation | 5.8 ± 3.3 | |
| Median (minimum; maximum) | 5 (1; 43) | |
| Days between hospital admission and surgery | ||
| Mean and standard deviation | 3.7 ± 6.4 | |
| Median (minimum; maximum) | 3 (- ; 127) | |
| Days between fracture and surgery | ||
| Mean and standard deviation | 12.4 ± 12.3 | |
| Median (minimum; maximum) | 10 (1; 135) | |
| Time between fracture and surgery > 48 h | ||
| No | 20 | 3.7 |
| Yes | 522 | 96.3 |
Conventional sign used: - Numerical data equal to zero not resulting from rounding.
Death registers were examined among the 542 patients who underwent surgery, and it was identified that 66 (12.2%) of them had died within one year of the procedure. For two patients, the death date was unknown (although certified). The most common causes of death were pulmonary (36.4%) and cardiac or cardiovascular events (18.2%). However, six patients did not have a cause of death described, and the other five were registered as "deaths without assistance". The only case of cancer, specifically uterine cancer, was a pre-existing condition prior to the hip surgery. In the only case where "death due to hip fracture" was registered, the patient had been indicated for a second surgery, but the procedure had been delayed due to difficulties in transportation to the hospital, as described in the medical record (Table 3).
Table 3 - Cause of death in all assessed patients, according to the ICD-10 code, registered in SESPA's database, 2015-2017
| Death cause / ICD-10 code | Days between surgery and death |
|---|---|
| Pulmonary events | |
| I26 - Pulmonary embolism | 324 |
| J96.0 - Acute respiratory failure | 245 |
| J96.0 - Acute respiratory failure | 71 |
| J96.0 - Acute respiratory failure | 138 |
| J96.0 - Acute respiratory failure | 66 |
| J96.0 - Acute respiratory failure | 156 |
| J96.0 - Acute respiratory failure | 125 |
| J96.0 - Acute respiratory failure | 36 |
| J96.0 - Acute respiratory failure | 14 |
| J96.0 - Acute respiratory failure | 6 |
| J96.0 - Acute respiratory failure | 132 |
| J96.0 - Acute respiratory failure | 129 |
| J96.0 - Acute respiratory failure | 30 |
| J96.0 - Acute respiratory failure | 53 |
| J96.0 - Acute respiratory failure | 1 |
| R09.2 - Respiratory arrest | 18 |
| R09.2 - Respiratory arrest | 23 |
| R09.2 - Respiratory arrest | 13 |
| R09.2 - Respiratory arrest | 312 |
| R09.2 - Respiratory arrest | 73 |
| R09.2 - Respiratory arrest | 41 |
| R09.2 - Respiratory arrest | 266 |
| R09.2 - Respiratory arrest | 187 |
| R09.2 - Respiratory arrest | 2 |
| Cardiovascular events | |
| I63.9 - Cerebral infarction, unspecified | 72 |
| I63.9 - Cerebral infarction, unspecified | 142 |
| I63.9 - Cerebral infarction, unspecified | 143 |
| I63.9 - Cerebral infarction, unspecified | 175 |
| I60.9 - Nontraumatic subarachnoid hemorrhage, unspecified | 4 |
| Cardiac events | |
| I48 - Atrial fibrillation and flutter | 14 |
| I21 - Acute myocardial infarction | 9 |
| I46 - Cardiac arrest | 12 |
| I46 - Cardiac arrest | 262 |
| I46 - Cardiac arrest | 101 |
| I46.9 - Cardiac arrest, cause unspecified | 254 |
| I46.9 - Cardiac arrest, cause unspecified | 318 |
| Shock | |
| R57.0 - Cardiogenic shock | 141 |
| R57.0 - Cardiogenic shock | 76 |
| R57.0 - Cardiogenic shock | 23 |
| R57.8 - Other shock | 110 |
| R57.8 - Other shock | 122 |
| R57.8 - Other shock | 357 |
| R57.9 - Shock, unspecified | 133 |
| Unspecified septicemia | |
| A41.9 - Sepsis, unspecified organism | 286 |
| A41.9 - Sepsis, unspecified organism | 258 |
| A41.9 - Sepsis, unspecified organism | 85 |
| A41.9 - Sepsis, unspecified organism | 64 |
| A41.9 - Sepsis, unspecified organism | 134 |
| A41.9 - Sepsis, unspecified organism | 343 |
| A41.9 - Sepsis, unspecified organism | 55 |
| A41.9 - Sepsis, unspecified organism | 7 |
| A41.9 - Sepsis, unspecified organism | 159 |
| Uterine cancer | |
| C55 - Malignant neoplasm of uterus, part unspecified | 268 |
| Femoral fracture | |
| S72 - Fracture of femur | 20 |
| Dementia | |
| F03 - Unspecified dementia | 48 |
| Not defined | |
| R99 - Ill-defined and unknown cause of mortality | 265 |
| R99 - Ill-defined and unknown cause of mortality | 158 |
| R99 - Ill-defined and unknown cause of mortality | 117 |
| R99 - Ill-defined and unknown cause of mortality | 89 |
| R99 - Ill-defined and unknown cause of mortality | 192 |
| R99 - Ill-defined and unknown cause of mortality | 4 |
| Death without assistance | |
| R98 - Unattended death | 41 |
| R98 - Unattended death | 309 |
| R98 - Unattended death | 17 |
| R98 - Unattended death | 32 |
| R98 - Unattended death | 324 |
In the univariate analysis, the variables identified as related to death in one year were age, sex, days between fracture and hospital admission, days between hospital admission and surgery, and days between fracture and surgery (Table 4). However, in the multivariate logistic regression (Table 5), the days between fracture and surgery (OR = 3.04; P = 0.169) are not significantly related to mortality in one year. Despite being the most common causes of death, respiratory comorbidities (OR = 3.04; P = 0.070) were not related to death in one year after hip surgery either. For every one year added to age, the risk of death increased 5.5%. Men with hip fractures had a 111% higher chance of dying one year after hip surgery than women.
Table 4 - A description of one-year mortality based on each evaluated characteristic and the results of the unadjusted analysis for patients undergoing surgical treatment for hip fractures at Maradei Hospital, Pará State, Brazil, 2015-2016
| Variables | Death in one year | 95% CI | P-value | |||||
|---|---|---|---|---|---|---|---|---|
| No | Yes | Odds ratio | Inferior | Superior | ||||
| N = 476 | % | N = 66 | % | |||||
| Age (interval) | ||||||||
| Mean and standard deviation | 79.5 ± 9.4 | 83.7 ± 9.5 | 1.049 | 1.019 | 1.078 | 0.001* | ||
| Age (years old) | ||||||||
| 60 to 69 | 81 | 92.0 | 7 | 8.0 | 1 | 0.105 | ||
| 70 to 79 | 138 | 90.8 | 14 | 9.2 | 1.17 | 0.46 | 3.03 | |
| 80 to 89 | 196 | 86.3 | 31 | 13.7 | 1.83 | 0.77 | 4.33 | |
| ≥ 90 | 61 | 81.3 | 14 | 18.7 | 2.66 | 1.01 | 6.98 | |
| Sex | ||||||||
| Female | 318 | 90.3 | 34 | 9.7 | 1 | 0.015 | ||
| Male | 158 | 83.2 | 32 | 16.8 | 1.89 | 1.13 | 3.18 | |
| Laterality | ||||||||
| Right | 225 | 88.9 | 28 | 11.1 | 1 | 0.460 | ||
| Left | 251 | 86.9 | 38 | 13.1 | 1.22 | 0.72 | 2.05 | |
| Controlled systemic arterial hypertension | ||||||||
| No | 266 | 86.9 | 40 | 13.1 | 1 | 0.468 | ||
| Yes | 210 | 89.0 | 26 | 11.0 | 0.82 | 0.49 | 1.39 | |
| Controlled diabetes mellitus | ||||||||
| No | 381 | 88.2 | 51 | 11.8 | 1 | 0.600 | ||
| Yes | 95 | 86.4 | 15 | 13.6 | 1.18 | 0.64 | 2.19 | |
| Stroke history | ||||||||
| No | 456 | 87.7 | 64 | 12.3 | 1 | > 0.999† | ||
| Yes | 20 | 90.9 | 2 | 9.1 | 0.71 | 0.16 | 3.12 | |
| Other cardiovascular disease | ||||||||
| No | 451 | 88.1 | 61 | 11.9 | 1 | 0.395† | ||
| Yes | 25 | 83.3 | 5 | 16.7 | 1.48 | 0.55 | 4.01 | |
| Alzheimer's disease | ||||||||
| No | 451 | 87.6 | 64 | 12.4 | 1 | 0.761† | ||
| Yes | 25 | 92.6 | 2 | 7.4 | 0.56 | 0.13 | 2.44 | |
| Pulmonary comorbidity | ||||||||
| No | 465 | 88.2 | 62 | 11.8 | 1 | 0.097† | ||
| Yes | 11 | 73.3 | 4 | 26.7 | 2.73 | 0.84 | 8.83 | |
| Smoker | ||||||||
| No | 405 | 87.9 | 56 | 12.1 | 1 | 0.960 | ||
| Yes | 71 | 87.7 | 10 | 12.3 | 1.02 | 0.5 | 2.09 | |
| Fracture site | ||||||||
| Femoral neck | 151 | 87.8 | 21 | 12.2 | 1 | 0.940‡ | ||
| Subtrochanteric | 24 | 85.7 | 4 | 14.3 | 1.2 | 0.38 | 3.8 | |
| Intertrochanteric | 301 | 88.0 | 41 | 12.0 | 0.98 | 0.56 | 1.72 | |
| Type of surgery | ||||||||
| Cephalomedullary nail | 225 | 86.5 | 35 | 13.5 | 1 | 0.807‡ | ||
| Bipolar hip hemiarthroplasty | 129 | 87.2 | 19 | 12.8 | 0.95 | 0.52 | 1.72 | |
| Total hip arthroplasty | 18 | 94.7 | 1 | 5.3 | 0.36 | 0.05 | 2.76 | |
| Dynamic hip screw plate | 81 | 90.0 | 9 | 10.0 | 0.71 | 0.33 | 1.55 | |
| External fixator | 2 | 100.0 | - | - | ND | |||
| Cannulated screws | 4 | 100.0 | - | - | ND | |||
| Dynamic condylar screw plate | 16 | 88.9 | 2 | 11.1 | 0.80 | 0.18 | 3.65 | |
| Girdlestone | 1 | 100.0 | - | - | ND | |||
| Days between fracture and hospital admission | 1.01 | 0.99 | 1.03 | |||||
| Mean and standard deviation | 8.4 ± 10.5 | 10.4 ± 12.3 | 0.008§ | |||||
| Median (minimum; maximum) | 5 (- ; 108) | 7 (- ; 92) | ||||||
| Days in hospital | ||||||||
| Mean and standard deviation | 5.7 ± 3.3 | 6.3 ± 3.3 | 1.05 | 0.98 | 1.11 | 0.078§ | ||
| Median (minimum; maximum) | 5 (1; 43) | 6 (3; 23) | ||||||
| Days between hospital admission and surgery | ||||||||
| Mean and standard deviation | 3.7 ± 6.7 | 4 ± 2.7 | 1.01 | 0,97 | 1.04 | 0.033§ | ||
| Median (minimum; maximum) | 3 (- ; 127) | 3.5 (1; 14) | ||||||
| Days between fracture and surgery | ||||||||
| Mean and standard deviation | 12.1 ± 12.3 | 14.4 ± 12.6 | 1.01 | 1.00 | 1.03 | 0.002§ | ||
| Median (minimum; maximum) | 9 (1; 135) | 12 (1; 95) | ||||||
| Time between fracture and surgery > 48 h | ||||||||
| No | 18 | 90.0 | 2 | 10.0 | 1 | > 0.999† | ||
| Yes | 458 | 87.7 | 64 | 12.3 | 1.26 | 0.29 | 5.55 | |
| City | ||||||||
| Belém (capital) | 193 | 86.9 | 29 | 13.1 | 1 | 0.599 | ||
| Countryside | 283 | 88.4 | 37 | 11.6 | 0.87 | 0.52 | 1.46 | |
Conventional sign used: - Numerical data equal to zero not resulting from rounding; CI: Confidence interval; p-values from Chi-square test except: * Student t-test; † Fisher's exact test; ‡ Likelihood ratio test; § Mann-Whitney test; ND: Unable to estimate.
Table 5 - Result of the multiple model to explain mortality in one year according to the evaluated characteristics of the patients undergoing hip fracture surgical treatment at Maradei Hospital, 2015-2016
| Variable | OR | 95% CI | P-value | |
|---|---|---|---|---|
| Inferior | Superior | |||
| Age (years old) | 1.053 | 1.023 | 1.084 | < 0.001 |
| Sex (male) | 2.11 | 1.24 | 3.60 | 0.006 |
| Pulmonary comorbidity | 3.04 | 0.92 | 10.06 | 0.07 |
| Days between fracture and surgery | 1.01 | 1.00 | 1.03 | 0.161 |
CI: Confidence interval; OR: Odds ratio.
DISCUSSION
This study identified a mortality rate of 12.2% in one year after hip surgery, with age and sex as the main variables associated with death. While the time between fracture and surgery was associated with mortality in isolated analysis, it lost statistical significance in multivariate analysis, a similar pattern with respiratory comorbidities. This result implies we could not substantiate the hypothesis that a prolonged interval between fracture and surgery impacts mortality. Nevertheless, the study highlights that elderly patients with hip fractures are succumbing to respiratory and cardiovascular diseases. We speculate that adequate primary care could have been prevented some of these deaths. Perhaps preventive measures to control high blood pressure, diabetes, and preventable infectious diseases are still finding barriers to reaching the communities that live alongside the rivers in the Amazon.
The death rate in this study is similar to the 12% rate found in a study in Taiwan, where the intertrochanteric fracture was more prevalent12. However, other studies have even higher rates, of 19% or 26%13,14.
The epidemiological profile of the patients analyzed here is not different from other national studies, which also show, for example, that the female sex is the most prevalent in hip fractures15,16. When analyzing one-year mortality, male gender is the most prevalent, consistent with findings in other studies6,17,18. Age affected the death rate: for each year added to age, there was a 5.5% higher chance of mortality in this study, a similar number to what was found in another evaluation in the South of Brazil, which registered an 8% increase per year in the risk19.
The hospital where this study was conducted is a referral center for hip surgery for the public system. It receives patients from SUS across the state that are not equipped or staffed to treat them surgically. Nearly 60% of our patients were non-residents in the capital, Belém, and required transportation by river and roads to reach the hospital for treatment. These trips took, on average, 8.4 days (from fracture to hospitalization). Inter-city transportation for hip surgery is common in Brazil20. Only 20 patients in this study underwent surgery within 48 h of the trauma. Cases not treated within this time window may entail a worse prognosis and higher healthcare costs associated with sequelae20,21; although, in fact, it has not affected the death rate in other studies22,23. We could not find any significant difference between these 20 and the remaining patients who had to wait more time.
Some authors argue that a delay exceeding 48 h may permit the stabilization of comorbidities that could otherwise impact surgical outcomes22. However, one must consider that the conditions of the hospitals where patients wait for surgery can be far from ideal, especially in the Amazon Region. The delay in surgery can affect prognosis and is associated with a higher risk of nosocomial pneumonia and reoperations due to other infections24,25. The higher hospital infection risk, especially among the elderly, speaks in favor of trying to operate in 48 h8,9.
Hip fractures mainly affect the elderly. Moreover, this population tends to present chronic comorbidities. It is natural that the death rates one year after hip fracture surgical treatment be related to diseases such as anemia, dementia, myocardial infarction, and chronic obstructive pulmonary disease5,22. The elderly with hip fractures tend to suffer from hypertension and diabetes, too16. However, no specific association between these chronic diseases and mortality after one year was found in this study sample.
The most frequently recorded causes of death among the patients within one year were pulmonary and cardiovascular events. The confidence in these data, sourced from death rate certificates within the Pará State system, might be appropriately questioned. However, the limitations in information quality and the absence of specific details on pulmonary diseases or cardiac events causing death hinder more in-depth analyses26,27,28.
Another limitation of the present study is the small sample. It is known that there is underreporting of deaths in Pará State because families in riverside communities fail to notify the death of relatives due to social and economic restraints. It is estimated that 7% of deaths in Pará are not notified29,30, and although we tried to mitigate this by making telephone calls to the families, we could not identify further deaths. Also, there is the possibility that some patients have died even before receiving adequate treatment, and so were not included in this research. Notwithstanding, the sample highly represents what happens in the Pará region and allows future comparisons.
CONCLUSION
Considering the local realities, ensuring that patients have access to surgical treatment for hip surgery 48 h after the trauma can be challenging. Although a significant association between this delay and mortality was not found, we conclude that reducing this time is justified by preventing complications that can impact patients' lives and healthcare systems. More extensive and in-depth investigations into the causes of death among patients operated for hip fractures are necessary to facilitate better planning of healthcare and logistics, particularly in the context of transporting patients in the Amazon.
