JOURNAL READING

Spinal Or General Anaesthesia For Surgical Repair Of Hip

Fracture And Subsequent Risk Of Mortality And Morbidity: A
Database Analysis Using Propensity Score-Matching

Pembimbing:
Dr. M. F. Susanti Handayani, Sp. An

Disusun Oleh :
Adeta Yuniza Mulia
2015730002

KEPANITERAAN KLINIK

STASE ILMU ANESTESI

RUMAH SAKIT UMUM DAERAH CIANJUR

FAKULTAS KEDOKTERAN DAN KESEHATAN

UNIVERSITAS MUHAMMADIYAH JAKARTA

2020
Anestesi Tulang Belakang Atau Umum Untuk Perbaikan Bedah Hip Fraktur
Dan Risiko Kematian Dan Morbiditas Selanjutnya: Analisis Basis Data
Menggunakan Pencocokan Skor Kecenderungan

Ringkasan
Sekitar 76.000 orang mengalami patah tulang pinggul mereka setiap tahun di
Inggris dengan biaya pribadi, sosial dan finansial yang besar. Meskipun
perdebatan telah berlangsung lama, mode anestesi yang optimal (umum atau
tulang belakang) masih belum jelas. Tujuan kami adalah untuk menilai apakah
ada perbedaan yang signifikan dalam mortalitas dan morbiditas antara pasien yang
menjalani anestesi spinal dibandingkan dengan anestesi umum selama operasi
fraktur panggul. Analisis sekunder memeriksa apakah ada perbedaan dalam
mortalitas untuk pasien dengan penyakit kardiovaskular yang sudah ada
sebelumnya atau penyakit paru obstruktif kronis. Ini adalah analisis database
klinis pasien yang dirawat karena patah tulang pinggul di Nottingham, UK antara
2004 dan 2015. Pencocokan skor kecenderungan digunakan untuk menghasilkan
pasangan yang cocok pasien, salah satunya menjalani masing-masing mode
anestesi. Data dianalisis menggunakan logistik bersyarat regresi, dengan 7164
pasien berhasil dicocokkan. Tidak ada perbedaan dalam mortalitas 30- atau 90
hari di Indonesia pasien yang memiliki anestesi spinal daripada anestesi umum
(OR [95% CI] 0,97 [0,8-1,15]; p = 0,764 dan 0,93 [0,82- 1.05]; p = masing-
masing 0,247). Pasien yang memiliki anestesi tulang belakang memiliki risiko
lebih rendah untuk transfusi darah (OR [95% CI] 0,84 [0,75-0,94]; p = 0,003) dan
infeksi saluran kemih (OR [95% CI] 0,72 [0,61-0,84]; p <0,001), tetapi lebih
mungkin mengembangkan infeksi dada (OR [95% CI] 1,23 [1,07-1,42]; p =
0,004), trombosis vena dalam (OR [95% CI] 2,18 [1,07-4,45]; p = 0,032) atau
emboli paru (OR [95% CI] 2,23 [1,16-4,29]; p = 0,016). Itu mode anestesi untuk
operasi fraktur panggul menghasilkan tidak ada perbedaan yang signifikan dalam
mortalitas, tetapi ada perbedaan signifikan dalam beberapa ukuran morbiditas
pasca operasi.

PENDAHULUAN
Patah tulang pinggul, atau fraktur sepertiga proksimal tulang paha [1], adalah
salah satu cedera serius paling umum yang terjadi pada populasi yang lebih tua.
Pada 2016, lebih dari 70.000 pasien berusia ≥ 60 tahun dirawat karena patah
tulang pinggul di rumah sakit di seluruh Inggris, biaya NHS dan perawatan sosial
£ 1 miliar (€ 1,13 miliar, US $ 1,26 miliar) setiap tahun [2]. Selanjutnya, satu
proyeksi menunjukkan bahwa pada tahun 2033, terlepas dari kejadiannya patah
tulang pinggul pada orang tua menurun, peningkatan populasi berisiko selama
periode intervensi berarti sekitar 100.000 pasien setiap tahun akan mengalami
patah tulang pinggul diperbaiki secara operasi di Inggris [3]. Akun populasi yang
lebih tua untuk sebagian besar patah tulang pinggul dan pasien ini menyajikan
pertimbangan anestesi tambahan dan keprihatinan; ini mungkin merupakan
cerminan dari pengurangan fungsional cadangan 'dalam sistem organ yang menua
[4]. Ini artinya tuntutan fisiologis tambahan, apakah itu intra atau perioperatif,
dapat menyebabkan penurunan yang cukup [5]. Ada sejumlah opsi anestesi untuk
pasien menjalani operasi fraktur panggul, tetapi tulang belakang atau umum
anestesi adalah dua yang paling sering digunakan di Inggris [6]. Salah satu
manfaat utama anestesi umum adalah kurangnya pengetahuan atau ingatan pasien
tentang prosedur, yang berpotensi meningkatkan kepuasan pasien, diberikan ruang
operasi yang tidak dikenal dan sering menakutkan lingkungan [7]. Namun, ada
manfaat potensial terkait dengan anestesi spinal, terutama kurang intraoperatif
hipotensi, penghindaran aktif secara neurologis obat-obatan dan kemungkinan
pengurangan delirium awal [8]. Keputusan mengenai anestesi diambil pada case-
bycase dasar; Institut Kesehatan dan Perawatan Nasional saat ini Pedoman
Excellence (NICE) menyatakan bahwa pasien seharusnya menawarkan pilihan
antara anestesi umum dan spinal mengikuti diskusi tentang manfaat masing -
masing dan kelemahan [9]. Dalam praktiknya, preferensi ahli anestesi mungkin
memainkan peran sentral dalam pengambilan keputusan ini. Penelitian yang ada
di bidang ini belum menunjukkan secara klinis perbedaan signifikan dalam
mortalitas dan morbiditas antara anestesi spinal dan umum untuk operasi fraktur
panggul. Di sebagian, ini mungkin karena masalah pengganggu oleh indikasi;
orang-orang yang berisiko lebih tinggi dari hasil yang merugikan mungkin lebih
mungkin untuk mendapatkan satu bentuk anestesi daripada yang lain. Pencocokan
skor kecenderungan adalah teknik statistik yang bertujuan untuk memberikan
analisis quasiexperimental di mana kelompok serupa di seluruh faktor pembaur
yang mungkin. Tujuan dari Penelitian observasional ini menggunakan
scorematching propensity untuk menyelidiki apakah ada secara klinis perbedaan
yang relevan dalam hasil antara tulang belakang dan umum anestesi di seluruh
populasi.
METODE
Data diperoleh dari Fraktur Pinggul Nottingham Basis data, registri klinis yang
berisi data tentang pra-penerimaan status kesehatan, intervensi bedah, dan pasca
bedah komplikasi untuk semua pasien yang telah menjalani operasi untuk
memperbaiki patah tulang pinggul di Universitas Nottingham Rumah Sakit NHS
Trust sejak 1999. Persetujuan untuk penggunaan dataset sepenuhnya anonim
untuk proyek ini diperoleh dari Tim Keamanan dan Risiko Kualitas Klinis
setempat. Sabar data yang dapat diidentifikasi (termasuk tanggal lahir, usia dan
tanggal masuk) dikeluarkan dari dataset untuk memastikan anonimitas.
Identifikasi semu disediakan untuk memungkinkan klarifikasi data antara tim
audit klinis dan para penyelidik. Pasien yang dirawat antara 2004 dan 2015
dimasukkan dalam penelitian ini. Pasien yang menjalani anestesi epidural tidak
dimasukkan. Itu sifat dan kualitas data ini telah dijelaskan dalam detail dalam
laporan sebelumnya [10]. Komorbiditas dan hasil dicatat berdasarkan panitera
masuk dan medis catatan. Paparan utama adalah dikotomi antara anestesi umum
dan spinal. Anestesi umum termasuk yang dengan dan tanpa blok saraf regional.
Demikian pula, anestesi spinal termasuk yang dengan dan tanpa blok saraf
regional. Ukuran hasil pasca operasi morbiditas direkam jika didokumentasikan
dalam catatan pasien oleh dokter yang merawat. Data mengenai mortalitas 30 hari
adalah diperoleh dan dirujuk silang dari data rumah sakit, juga data dari Kantor
Statistik Nasional (ONS), seperti yang telah telah dijelaskan sebelumnya [10].
Rumah sakit rujukan silang data dengan data ONS memungkinkan identifikasi
semua pasien yang telah meninggal di masyarakat, atau di rumah sakit lain.
Ukuran hasil primer adalah mortalitas 30 hari. Hasil sekunder yang
mendefinisikan apriori adalah: 90 hari kematian; persyaratan untuk transfusi
darah; pasca operasi infeksi (infeksi luka dalam, infeksi saluran kemih (ISK) dan
infeksi Clostridium difficile); gagal jantung; trombosis vena dalam (DVT); emboli
paru (PE); infark miokard; kecelakaan serebrovaskular; luka hematoma; gagal
ginjal; dan perdarahan gastrointestinal. Dua sub-kelompok non-eksklusif
didefinisikan a apriori berdasarkan risiko yang diketahui lebih tinggi dari hasil
yang merugikan. Ini adalah pasien dengan kardiovaskular yang terdokumentasi
penyakit (kondisi kardiovaskular yang sudah ada termasuk: infark miokard
sebelumnya; penyakit jantung iskemik; fibrilasi atrium; penyakit jantung katup;
atau hipertensi) dan penyakit paru obstruktif kronik (PPOK) [11]. Data mentah
dibersihkan sebelum analisis. Jelas nilai yang salah diperiksa kembali dengan data
asli dan diperbaiki. Data yang hilang tidak dapat diperbaiki diberi kode sebagai
variabel dummy missing. Melepaskan lokasi dikodekan ulang menjadi: rumah
sakit; rumah tinggal; sendiri rumah; atau yang lain. Pencocokan skor
kecenderungan digunakan untuk mensimulasikan atribut uji coba terkontrol secara
acak dalam suatu desain penelitian observasional. Delapan belas kovariat berbeda
digunakan dalam model skor kecenderungan: tahun penerimaan; usia saat masuk
(banded ke 5 tahun); seks; jenis fraktur, Status fisik ASA; apakah pasien
memerlukan bantuan untuk berjalan sebelum masuk; skor tes mental yang
disingkat [12]; adanya penyakit kardiovaskular; sebelumnya kecelakaan
serebrovaskular; diagnosis COPD; kehadiran dari penyakit ginjal; diagnosis
diabetes; keganasan yang ada; status merokok; polifarmasi (≥ empat obat
ditentukan); konsentrasi hemoglobin saat masuk; tingkat ahli bedah yang
beroperasi; dan tingkat kepedulian ahli anestesi. Pasangan yang cocok dibentuk
dengan menggunakan terdekat pencocokan tetangga, meminimalkan perbedaan
antara skor kecenderungan pasien berpasangan [13]. Kecenderungan pencocokan
skor dilakukan secara terpisah untuk subkelompok pasien kardiovaskular dan
PPOK. Bersyarat regresi logistik kemudian dilakukan, membandingkan tulang
belakang anestesi dengan anestesi umum. Semua analisis adalah dilakukan dengan
menggunakan paket perangkat lunak statistik Stata / SE 15.1 (StataCorp LLC,
College Station, TX, USA).
HASIL
Secara total, 8144 pasien dilibatkan dalam analisis awal, semuanya di antaranya
memiliki skor kecenderungan yang dihitung (Gbr. 1). Dari ini, 6054 (74,3%)
adalah perempuan, dan yang paling umum braket usia saat masuk adalah 85-89
tahun (24,6%). Dalam hal kondisi komorbiditas, 4.965 pasien memiliki
kardiovaskular penyakit (61,0%) dan 1391 (17,1%) menderita COPD. Selama
operasi, 1312 (16,1%) pasien menjalani operasi mereka dilakukan oleh konsultan
bedah, sedangkan 5253 (64,5%) punya konsultan anestesi. Prevalensi beragam
ukuran hasil dalam anestesi umum dan spinal kohort dirinci dalam Tabel 1.
Pencocokan terdekat-tetangga berusaha untuk mencocokkan pasien dengan mitra
yang cocok, dan 7.164 pasien dicocokkan menjadi 3.582 pasangan. Tidak ada
perbedaan signifikan dalam 30- atau 90 hari mortalitas pada pasien yang memiliki
spinal daripada umum anestesi (OR [95% CI] 0,97 [0,8-1,15]; p = 0,764 dan 0,93
[0,82–1,05]; p = 0,247) masing-masing (ATAU <1 nikmat anestesi spinal).
Mengingat bahwa pedoman nasional, protokol manajemen dan faktor sistematis
lainnya berubah seiring waktu, kami mengulangi analisis hanya menggunakan
72,8% pasangan di mana operasi masing-masing dalam 3 tahun lain. Sekali lagi,
tidak ada perbedaan signifikan dalam 30- atau 90- mortalitas hari pada pasien
yang memiliki spinal daripada umum anestesi (OR [CI 95%] 0,92 [0,74-1,15]; p =
0,460 dan 0,95 [0,81-1,13]; p = 0,580) masing-masing. Dampak spinal
dibandingkan dengan umum anestesi pada kejadian hasil sekunder tindakan
ditunjukkan pada Tabel 2. Anestesi spinal ditemukan menjadi pelindung untuk
dua faktor: darah pasca operasi transfusi (OR [95% CI] 0,84 [0,75-0,94]; p =
0,003); dan ISK pasca operasi (OR [95% CI] 0,72 [0,61-0,84]; p <0,001). Namun,
pasien yang menerima anestesi spinal adalah lebih mungkin mengembangkan
infeksi dada pasca operasi (OR) [95% CI] 1,23 [1,07-1,42]; p = 0,004), emboli
paru (PE) (ATAU [95% CI] 2,23 [1,16-4,29]; p = 0,016) atau vena dalam
trombosis (DVT) (OR [95% CI] 2,18 [1,07-4,45]; p = 0,032). Tidak ada ukuran
lain dari morbiditas pasca operasi yang menunjukkan a perbedaan yang signifikan
secara statistik. Penggunaan anestesi spinal tidak mempengaruhi 30 hari
mortalitas pada pasien dengan penyakit kardiovaskular (OR [95% CI] 0,91 [0,74-
1,12]; p = 0,372) atau COPD (OR [95% CI] 0,98 [0,66-1,45]; p = 0,920). Namun,
anestesi spinal dikaitkan dengan penurunan yang signifikan secara statistik pada
90- kematian hari pada pasien dengan penyakit kardiovaskular (OR [95% CI] 0,84
[0,72-0,98]; p = 0,026) tetapi tidak untuk mereka yang COPD (ATAU [95% CI]
0,92 [0,68-1,24]; p = 0,590).
DISKUSI
Dalam studi ini, mode anestesi tidak mempengaruhi 30- atau Mortalitas 90 hari
pada populasi umum pasien memiliki patah tulang pinggul yang diperbaiki secara
operasi. Pasien-pasien itu menerima anestesi spinal cenderung membutuhkan
transfusi darah pasca operasi atau mengembangkan ISK. Namun, mereka berada
pada risiko yang meningkat untuk mengembangkan pasca operasi infeksi dada
atau tromboemboli vena. Tidak ada bukti yang meyakinkan untuk dampak dari
mode anestesi pada kematian untuk sub-kelompok pasien yang ditentukan
sebelumnya dengan penyakit kardiovaskular atau COPD. Temuan ini berkaitan
dengan kematian sejalan dengan bukti saat ini. Tinjauan sistematis pasien dewasa
dengan patah tulang pinggul tidak menemukan perbedaan dalam mortalitas 30
hari dengan anestesi umum vs. regional, berdasarkan 11 penelitian
menggabungkan 2152 peserta [14]. Sistematik yang sama Ulasan juga tidak
menemukan perbedaan dalam mortalitas pada 3 bulan berdasarkan lima studi dan
953 peserta [14]. Ini adalah sesuai dengan tinjauan sistematis lain termasuk 14
penelitian [15] dan penelitian besar di Amerika (> 50.000 pasien) [16] bahwa
keduanya menunjukkan bahwa teknik anestesi tidak punya efek pada mortalitas
30 hari.
Data hasil sekunder adalah pembuatan hipotesis dan harus ditafsirkan dengan hati-
hati, khususnya analisis sub-kelompok pasien dengan penyakit kardiovaskular dan
COPD. Terjadi pengurangan pada transfusi darah pasca operasi setelah anestesi
spinal. Ada sejumlah penjelasan potensial untuk ini temuan. Mungkin saja
hubungan antara tulang belakang benar anestesi dan kebutuhan untuk darah pasca
operasi transfusi ada. Sebuah penelitian menemukan bahwa pasien yang menjalani
anestesi spinal untuk artroplasti panggul 35% lebih kecil kemungkinannya
membutuhkan transfusi pasca operasi, dan penulis mendalilkan bahwa anestesi
spinal menyebabkan intraoperatif yang lebih rendah tekanan darah dan karenanya
pengurangan kehilangan darah [17]. Namun, temuan obat bius audit praktik sprint
menyarankan agar intra-operatif tekanan darah arteri lebih rendah dengan anestesi
umum pada pasien yang menjalani operasi fraktur panggul [6]. Selain itu, kami
hasilnya sesuai dengan meta-analisis dari 66 uji coba terkontrol secara acak yang
menemukan perbedaan rata-rata dari perkiraan kehilangan darah menjadi 335 ml
lebih rendah dengan tulang belakang dibandingkan dengan anestesi umum dalam
berbagai operasi prosedur [18]. Ada kemungkinan bahwa tingkat residual
pengganggu mungkin menyebabkan temuan ini karena tidak ada data tersedia
untuk mencocokkan pasien berdasarkan koagulopati yang ada; kondisi ini
merupakan kontra-indikasi relatif terhadap tulang belakang anestesi, karena
peningkatan risiko kanalis vertebral hematoma dan kompresi sumsum tulang
belakang berikutnya [19]. Mengingat bahwa pasien dengan koagulopati yang ada
memiliki peningkatan kemungkinan membutuhkan pasca operasi transfusi [20],
dan juga lebih cenderung dianggap tidak cocok untuk anestesi spinal, mungkin
saja ini terjadi efek perancu. Asosiasi anestesi spinal dengan pengurangan ISK
pasca operasi mengejutkan. Ada sangat sedikit bukti dalam literatur yang secara
langsung mendukung atau bertentangan dengan temuan ini, karena penelitian
sebelumnya cenderung fokus pada retensi urin pasca operasi (yang mungkin
selanjutnya membutuhkan kateterisasi dan meningkatkan risiko pasien yang
mengalami ISK). Beberapa penelitian ini memiliki menyarankan bahwa anestesi
spinal meningkatkan kemih retensi dibandingkan dengan anestesi umum. Namun,
sebagian besar kateter urin di unit kami ditempatkan sebelum operasi sesuai
dengan protokol departemen. Penelitian di PT daerah ini tidak konsisten [21, 22]
dan efek relatif yang sebenarnya anestesi spinal dan umum pada retensi urin dan
ISK tidak jelas. Mungkin saja tingkat kateterisasi dapat terjadi telah mengubah
rekaman klinis diagnosis ISK, yang terkenal tidak bisa diandalkan. Tes dipstick
positif dalam a pasien yang dikateterisasi mungkin dianggap kurang pentingnya
daripada jika pasien tidak dikateterisasi.
Kami menemukan bahwa pasien yang menerima spinal anestesi secara signifikan
lebih mungkin menderita DVT atau PE pasca operasi. Tinjauan sistematis tidak
menemukan perbedaan signifikan dalam kejadian DVT ketika chemoprophylaxis
digunakan [14], sebuah praktik yang sekarang lumrah [23] dan rutin di lembaga
kami; namun, penelitian itu menunjukkan bahwa anestesi spinal meningkatkan
risiko PE pasca operasi meskipun ini tergantung pada jenis analisis statistik yang
dilakukan [14]. Itu mungkin bahwa pasien yang memiliki risiko VTE lebih besar
(karena risiko dasar atau komorbiditas yang mengurangi mobilisasi) lebih
mungkin diberikan anestesi spinal. Sana oleh karena itu, dapat menjadi dukungan
dalam bukti saat ini untuk hubungan antara anestesi spinal dan peningkatan risiko
kejadian tromboemboli pada fraktur panggul operasi. Peningkatan infeksi dada
pasca operasi di dalam kelompok anestesi spinal mungkin merupakan
konsekuensi dari dua faktor yang berlawanan. Anestesi spinal umumnya diyakini
sebagai pelindung atau netral untuk pernapasan komplikasi [14, 15] tetapi ini pada
gilirannya dapat menyebabkan lebih besar penggunaan anestesi spinal pada pasien
yang berisiko pernapasan komplikasi [24]. Proporsi pasien dengan COPD yang
menerima anestesi spinal hampir dua kali lipat dari mereka yang memiliki anestesi
umum; ini mungkin mencerminkan pengambilan keputusan klinis yang disengaja
untuk menghindari umum anestesi pada pasien dengan penyakit pernapasan yang
sudah ada sebelumnya. Namun, ini memperkenalkan kemungkinan baseline
membingungkan; pasien-pasien dengan COPD yang ada adalah lebih mungkin
untuk mengembangkan infeksi dada pasca operasi dan relatif terwakili dalam
populasi yang memiliki anestesi tulang belakang. Argumen klinis dibuat sering
bahwa pasien dengan patologi dasar yang spesifik, khususnya COPD, lebih baik
dilayani oleh anestesi spinal. Kami tidak dapat menemukan bukti yang
meyakinkan tentang manfaat ini. Seperti halnya semua studi non-acak, kami tidak
dapat mengecualikan residu perancu, khususnya mengenai tingkat keparahan
COPD. Ada batasan untuk penelitian ini yang seharusnya diakui. Karakteristik
klinis pra-operasi dan morbiditas setelah operasi dicatat tidak selalu didefinisikan
dengan jelas. Karena itu, ada potensi untuk bias kesalahan klasifikasi jika data
direkam dengan cara yang berbeda sebagai akibat dari perbedaan interpretasi dan /
atau diagnosis antara dokter individu. Jangka waktu di mana data dicatat ke dalam
basis data dan perubahan dalam praktek klinis selama periode waktu tersebut akan
dengan sendirinya berpotensi mempengaruhi konkordansi berbagai dokter '
pendapat, apalagi variasi yang melekat yang akan ada di antara praktisi terlepas.
Selain itu, meskipun metode penyesuaian untuk membatasi efek perancu, itu
adalah mungkin bahwa itu sebenarnya telah meningkatkan risiko
mencampuradukkan, sebagai akibat dari efek yang digambarkan sebagai ‘the
paradoks skor kecenderungan '[25, 26]. Dalam konteks ini studi, di mana kovariat
sudah seimbang di antara kedua kelompok, ‘paradoks’ menjelaskan cara teknik
pencocokan skor kecenderungan mungkin telah menyebabkan ketidakseimbangan
dalam distribusi klinis karakteristik; ini terjadi melalui pemangkasan pasangan
dengan perbedaan terbesar dalam skor kecenderungan, dan berpotensi
memperkenalkan bias ke dalam analisis [26]. Selain itu, dataset tidak termasuk
informasi tentang tingkat delirium atau perubahan fungsi kognitif. Meskipun data
dari kualitas terbatas, ini mungkin di mana manfaat tulang belakang kebohongan
anestesi (jika ada yang bisa ditemukan). Lain Keterbatasan adalah bahwa anestesi
spinal dan umum telah terjadi dianggap sebagai entitas tunggal; kami tidak dapat
mempertimbangkan bagaimana ini disampaikan. Penelitian mungkin
membandingkan 'anestesi spinal' buruk 'dengan' baik 'umum anestesi atau
sebaliknya. Demikian pula, kita tidak bisa mengomentari dampak sedasi yang
digunakan bersama dengan anestesi spinal [27, 28]. Hingga pengumpulan data
rutin memungkinkan analisis tentang bagaimana anestesi disampaikan, ini
pertanyaan cenderung tetap tidak terjawab. Akhirnya, tapi mungkin yang paling
penting, kita mungkin hanya mengukur hasil yang salah [29]. Jika hasil klinis
tidak terpengaruh oleh cara anestesi, kemudian menangani hasil yang lebih lembut
seperti kepuasan pasien, kualitas hidup dan operasi efisiensi teater mungkin lebih
penting. Terlepas dari keterbatasan ini, ada kekuatan utama: dengan kekuatan
dataset 8000-pasien, dan data tentang amultitude hasil, kami tidak dapat
menunjukkan relevan secara klinis perbedaan antara mode anestesi. Untuk dokter
individu menyatakan bahwa tekniknya disukai optimal berdasarkan data tidak
lengkap yang hanya melibatkan puluhan atau ratusan pasien adalah klaim yang
tidak dapat dibuktikan. Ada empat uji coba terkontrol acak yang sedang
berlangsung dari anestesi spinal vs umum [30-33]. Kami menunggu hasilnya ini
dengan bunga. Sementara itu, tampaknya keputusan untuk menggunakan mode
anestesi tertentu untuk a pasien yang menjalani operasi fraktur panggul terutama
didasarkan sekitar pilihan pasien dan kecenderungan individu praktisi [34],
daripada pendekatan standar. Kita menyarankan bahwa penelitian masa depan
harus diarahkan bagaimana anestesi disampaikan, bukan mode anestesi per se.
 Original Article

Spinal or general anaesthesia for surgical

repair of hip fracture and subsequent risk
of mortality and morbidity: a database
analysis using propensity score-matching*
L. Morgan,1 T. M. McKeever,2 J.Nightingale,3 D. E. Deakin4 and I. K.
Moppett5

1 Medical Student, 2 Professor, Division of Epidemiology and Public Health, 5 Professor,

Anaesthesia and Critical Care, Division of Clinical Neuroscience,
University of Nottingham, UK 3 Research Manager, 4 Consultant,
Department of Trauma and Orthopaedics, Nottingham University Hospitals NHS Trust,
Nottingham, UK

Summary
Around 76,000 people fracture their hip annually in the UK at a considerable personal, social and financial
cost. Despite longstanding debate, the optimal mode of anaesthesia (general or spinal) remains unclear. Our
aim was to assess whether there is a significant difference in mortality and morbidity between patients
undergoing spinal anaesthesia compared with general anaesthesia during hip fracture surgery. A secondary
analysis examined whether a difference exists in mortality for patients with pre-existing cardiovascular
disease or chronic obstructive pulmonary disease. This was a clinical database analysis of patients treated for
hip fracture in Nottingham, UK between 2004 and 2015. Propensity score-matching was used to generate
matched pairs of patients, one of whom underwent each mode of anaesthesia. Data were analysed using
conditional logistic regression, with 7164 patients successfully matched. There was no difference in 30- or 90-
day mortality in patients who had spinal rather than general anaesthesia (OR [95%CI] 0.97 [0.8–1.15]; p =
0.764 and 0.93 [0.82– 1.05]; p = 0.247 respectively). Patients who had a spinal anaesthetic had a lower-risk of
blood transfusion (OR [95%CI] 0.84 [0.75–0.94]; p = 0.003) and urinary tract infection (OR [95%CI] 0.72 [0.61–
0.84]; p < 0.001), but were more likely to develop a chest infection (OR [95%CI] 1.23 [1.07 –1.42]; p = 0.004),
deep vein thrombosis (OR [95%CI] 2.18 [1.07–4.45]; p = 0.032) or pulmonary embolism (OR [95%CI] 2.23
[1.16–4.29]; p = 0.016). The
mode of anaesthesia for hip fracture surgery resulted in no signi ficant difference in mortality, but there was a
significant difference in several measures of postoperative morbidity.
................................................................................................................................................................
.
Corresponden
ce to: I.
Moppett
Email:
iain.moppett
@nottingham.
ac.uk
Accepted: 9
March 2020
Keywords: anaesthesia, general; anaesthesia, spinal; hip fracture; mortality; transfusion
*Presented in part at the Royal Medical Society’s National Student Conference,
Edinburgh, UK, January 2019. Twitter: @iainmoppett

Introduction (€1.13 billion, US$1.26 billion) annually [2].

Hip fractures, or fractures of the proximal
Furthermore, one projection indicates that by
third of the femur [1], are one of the most
2033, despite the incidence of hip fractures in
common serious injuries that occur in the
older people decreasing, the increase in the at
older population. In 2016, more than 70,000
risk population during the intervening period
patients aged ≥ 60 years were treated for a
means around 100,000 patients annually will
hip fracture in hospitals around the UK,
have a hip fracture fixed surgically in England
costing the NHS and social care £1 billion
[3]. The older population account

© 2020 The Authors. Anaesthesia published by John Wiley & Sons Ltd on behalf of Association of Anaesthetists
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial
License, which permits use, distribution and reproduction in any medium, provided the original work is
properly cited and is not used for commercial purposes.
Anaesthesia 2020 Morgan et al. | Effects of anaesthetic technique in hip fracture surgery
for the vast majority of hip fractures and these patients
present additional anaesthetic considerations and
concerns; this may be a reflection of the reduced
‘functional reserve’ in ageing organ systems [4]. This
means that additional physiological demands, whether
they be intra- or peri-operative, can lead to considerable
impairment [5].
There are a number of anaesthetic options for patients
having hip fracture surgery, but spinal or general
anaesthesia are the two used most commonly in the UK
[6]. One of the main benefits of a general anaesthetic is the
patient’s lack of knowledge or memory of the procedure,
which could potentially increase patient satisfaction, given
the unfamiliar and often daunting operating theatre
environment [7]. However, there are potential benefits
associated with spinal anaesthesia, notably less intra-
operative hypotension, avoidance of neurologically active
drugs and a possible reduction in early delirium [8].
Decisions regarding anaesthesia are taken on a case-by-
case basis; current National Institute of Health and Care
Excellence (NICE) guidelines state that patients should be
offered the choice between general and spinal anaesthesia
following a discussion on the respective benefits and
drawbacks [9]. In practice, anaesthetist preference
probably plays a central role in this decision-making.
Existing research in this area has not shown clinically
significant differences in mortality and morbidity between
spinal and general anaesthesia for hip fracture surgery. In
part, this may be due to the problem of confounding by
indication; people at higher risk of adverse outcomes may
be more likely to get one form of anaesthesia than
another. Propensity score-matching is a statistical
technique which aims to provide a quasi-experimental
analysis where groups are similar across possible
confounding factors. The aim of this observational study
was to use propensity score- matching to investigate
whether there was a clinically relevant difference in
outcomes between spinal and general anaesthesia across
the population.
Methods
The data were obtained from the Nottingham Hip Fracture
Database, a clinical registry that contains data on pre-
admission health status, surgical intervention and post-
discharge complications for all patients who have
undergone surgery to repair a hip fracture in Nottingham
University Hospitals NHS Trust since 1999. Approval for the
use of the fully anonymised dataset for this project was
gained from the local Clinical Quality Risk and Safety Team.
Patient identifiable data (including date of birth, age and
date of admission) were excluded from the dataset to
ensure anonymity. A pseudo-identifier was provided in
order to
allow clarification of data between the clinical audit team
and the investigators. Patients who were admitted
between 2004
and 2015 were included in the study. Patients who
underwent epidural anaesthesia were not included. The
nature and quality of these data have been described in
detail in previous reports [10]. Comorbidities and outcomes
are recorded based on the admission clerking and medical
records.
The primary exposure was a dichotomy between
general and spinal anaesthesia. General anaesthesia
included those with and without regional nerve blocks.
Similarly, spinal anaesthesia included those with and
without regional nerve blocks.
The postoperative outcome measures of morbidity
were recorded if documented in the patient’s notes by the
treating clinician. Data regarding 30-day mortality are
obtained and cross-referenced from hospital data, as well as
data from the Office for National Statistics (ONS), as has
been described previously [10]. Cross referencing hospital
data with ONS data allows identification of all patients who
have died in the community, or in other hospitals.
The primary outcome measure was 30-day mortality.
Secondary outcomes defined a priori were: 90-day
mortality; requirement for blood transfusion; postoperative
infections (deep wound infection, urinary tract infection
(UTI) and Clostridium difficile infection); cardiac failure;
deep vein thrombosis (DVT); pulmonary embolism (PE);
myocardial infarction; cerebrovascular accident; wound
haematoma; renal failure; and gastro-intestinal haem-
orrhage. Two non-exclusive sub-groups were defined a
priori based on a known higher risk of adverse outcomes.
These were patients with documented cardiovascular
disease (pre-existing cardiovascular conditions including:
previous myocardial infarction; ischaemic heart disease;
atrial fibrillation; valvular heart disease; or hypertension)
and chronic obstructive pulmonary disease (COPD) [11].
Raw data were cleaned before analysis. Clearly
erroneous values were back-checked with the original data
and corrected. Missing data that could not be corrected
were coded as a dummy missing variable. Discharge
location was recoded into: hospital; residential home; own
home; or other.
Propensity score-matching was used to simulate
attributes of a randomised controlled trial within an
observational study design. Eighteen different covariates
were used in the propensity score model: year of admission;
age at admission (banded to 5 years); sex; fracture type,
ASA physical status; whether the patient required help to
walk before admission; abbreviated mental test score [12];
presence of cardiovascular disease; previous
2 © 2020 The Authors.
Anaesthesia published by John
Anaesthesia 2020 Morgan et al. | Effects of anaesthetic technique in hip fracture surgery
Wiley & Sons Ltd on behalf
of Association of
Anaesthetists
 Morgan et al. | Effects of anaesthetic technique in hip fracture surgery Anaesthesia 2020

Figure 1 Flowchart to indicate inclusion/exclusion throughout the study.

cerebrovascular accident; diagnosis of COPD; presence of

renal disease; diagnosis of diabetes; existing malignancy;
smoking status; polypharmacy (≥ four medications
prescribed); haemoglobin concentration on admission;
grade of operating surgeon; and grade of
anaesthetist. Matched pairs were formed by using nearest
neighbour matching, minimising the difference between
the propensity scores of the paired patients [13].
Propensity score-matching was performed separately for
the
cardiovascular and COPD patient sub-groups. Conditional
logistic regression was then carried out, comparing spinal
anaesthesia with general anaesthesia. All analyses were
carried out using the statistical software package Stata/SE
15.1 (StataCorp LLC, College Station, TX, USA).
Results
In total, 8144 patients were included in the initial analysis,
all of whom had a propensity score calculated (Fig. 1). Of
these, 6054 (74.3%) were women, and the most common
age bracket on admission was 85–89 years (24.6%). In
terms of comorbid conditions, 4965 patients
cardiovascular disease (61.0%) and 1391 (17.1%) had
COPD. During surgery, 1312 (16.1%) patients had their
operation performed by a consultant surgeon, whereas
5253 (64.5%) had a consultant anaesthetist.
prevalence of the various outcome measures in the
general and spinal anaesthetic cohorts is detailed in Table
1. Nearest-neighbour matching attempted to match
patients with a suitable counterpart, and 7164 patients
were matched into 3582 pairs.
There was no significant difference in 30- or 90-day
mortality in patients who had spinal rather than general
anaesthesia (OR [95%CI] 0.97 [0.8–1.15]; p = 0.764 and
0.93 [0.82–1.05]; p = 0.247) respectively (OR < 1 favours
spinal anaesthesia). Given that national guidance,
management protocols and other systematic factors
caring change over time, we repeated the analysis using only the
72.8% of pairs where operations were within 3 years of
each other. Again, there was no significant difference in 30-
or 90- day mortality in patients who had spinal rather than
general anaesthesia (OR [95%CI] 0.92 [0.74–1.15]; p =
0.460 and
0.95 [0.81–1.13]; p = 0.580) respectively.
The impact of spinal compared with general
anaesthesia on the incidence of the secondary outcome
measures is shown in Table 2. Spinal anaesthesia was
found to be protective for two factors: postoperative blood
transfusion (OR [95%CI] 0.84 [0.75–0.94]; p = 0.003); and
postoperative UTI (OR [95%CI] 0.72 [0.61–0.84]; p < 0.001).
However, those patients receiving spinal anaesthesia were
more likely to develop a postoperative chest infection (OR
[95%CI] 1.23 [1.07–1.42]; p = 0.004), pulmonary embolism
(PE) (OR [95%CI] 2.23 [1.16–4.29]; p = 0.016) or deep vein
had thrombosis (DVT) (OR [95%CI] 2.18 [1.07–4.45]; p = 0.032).
No other measures of postoperative morbidity showed a
statistically significant difference.
The use of spinal anaesthesia did not affect 30-day
The mortality in patients with either cardiovascular disease (OR
[95%CI] 0.91 [0.74–1.12]; p = 0.372) or COPD (OR [95%CI]
0.98 [0.66–1.45]; p = 0.920). However, spinal anaesthesia
was associated with a statistically significant reduction in
90- day mortality in patients with cardiovascular disease
(OR [95%CI] 0.84 [0.72–0.98]; p = 0.026) but not for those
with COPD (OR [95%CI] 0.92 [0.68–1.24]; p = 0.590).
Morgan et al. | Effects of anaesthetic technique in hip fracture surgery Anaesthesia 2020
© 2020 The Authors. Anaesthesia published by John Wiley & Sons Ltd on behalf of Association of Anaesthetists 3
Anaesthesia 2020 Morgan et al. | Effects of anaesthetic technique in hip fracture surgery

Table 1 Comparison of the prevalence of postoperative outcome measures stratified by method of anaesthesia for surgical
repair of a hip fracture.
General anaesthesia Spinal anaesthesia
Outcome measure n = 4186 n = 3958
Blood transfusion 903 (21.6%) 739 (18.7%)
Chest infection 468 (11.2%) 523 (13.2%)
Urinary tract infection 474 (11.3%) 338 (8.5%)
Renal failure 296 (7.1%) 240 (6.1%)
Myocardial infarction 90 (2.2%) 89 (2.2%)
Cardiac failure 77 (1.8%) 58 (1.5%)
Haematoma 67 (1.6%) 48 (1.2%)
Clostridium difficile infection 39 (0.9%) 28 (0.7%)
Deep infection 39 (0.9%) 26 (0.7%)
Gastro-intestinal haemorrhage 32 (0.8%) 28 (0.7%)
Cerebrovascular accident 20 (0.5%) 23 (0.6%)
Pulmonary embolism 18 (0.4%) 29 (0.7%)
Deep vein thrombosis 12 (0.3%) 24 (0.6%)

Table 2 Comparison of postoperative morbidity for spinal anaesthesia vs. general anaesthesia for patients having surgical repair
of a hip fracture. Odds ratio < 1 favours spinal anaesthesia.
Outcome measure Discordant pairs Odds ratio (95%CI) p value
Blood transfusion 2306 0.84 (0.75–0.94) 0.003
Chest infection 1540 1.23 (1.07–1.42) 0.004
Urinary tract infection 1306 0.72 (0.61–0.84) < 0.001
Renal failure 844 0.83 (0.68–1.00) 0.052
Myocardial infarction 318 0.99 (0.72–1.35) 0.937
Cardiac failure 230 0.77 (0.53–1.11) 0.163
Haematoma 194 0.76 (0.51–1.14) 0.188
Deep infection 118 0.69 (0.41–1.15) 0.155
Clostridium difficile infection 114 0.73 (0.43–1.23) 0.235
Gastro-intestinal haemorrhage 104 0.93 (0.54–1.60) 0.782
Pulmonary embolism 84 2.23 (1.16–4.29) 0.016
Cerebrovascular accident 72 1.25 (0.65–2.41) 0.506
Deep vein thrombosis 70 2.18 (1.07–4.45) 0.032

Discussion These findings with regards to mortality are in line

In this study, the mode of anaesthesia did not affect 30- or
90-day mortality in the general population of patients
having a hip fracture surgically repaired. Those patients
receiving spinal anaesthesia were less likely to require a
postoperative blood transfusion or develop a UTI.
However, they were at increased risk of developing a
postoperative chest infection or
thromboembolism. There was no convincing evidence for
an impact of mode of anaesthesia on mortality for the pre-
specified sub-groups of patients with cardiovascular
disease or COPD.
with current evidence. A systematic review of adult
patients with hip fractures found no difference in 30-day
mortality with regional vs. general anaesthesia, based on
11 studies incorporating 2152 participants [14]. The same
systematic review also found no difference in mortality at 3
months based on five studies and 953 participants [14].
venous This is concordant with another systematic review
including 14 studies [15] and a large American study (>
50,000 patients)
[16] that both showed that anaesthetic technique had no
effect on 30-day mortality.

4 © 2020 The Authors. Anaesthesia published by John Wiley & Sons Ltd on behalf of Association of Anaesthetists
Morgan et al. | Effects of anaesthetic technique in hip fracture surgery Anaesthesia 2020

The secondary outcome data are hypothesis-

We found that patients who received spinal
generating and should be interpreted with caution,
anaesthesia were significantly more likely to suffer from
particularly the sub-group analyses of patients with
DVT or PE postoperatively. A systematic review found no
cardiovascular disease and COPD. There was a reduction in
significant difference in the incidence of DVT when
postoperative blood transfusions after spinal anaesthesia.
chemoprophylaxis was used [14], a practice which is now
There are a number of potential explanations for this
commonplace [23] and routine in our institution; however,
finding. It is possible a true relationship between spinal
the study did show that spinal anaesthesia increased the
anaesthesia and requirement for postoperative blood
risk of postoperative PE although this was dependent on
transfusion exists. A study found that patients undergoing
the type of statistical analysis performed [14]. It is possible
spinal anaesthesia for hip arthroplasty were 35% less likely
that patients who were at greater risk of VTE (due to
to require a postoperative transfusion, and the authors
baseline risk or comorbidities reducing mobilisation)
postulated that spinal anaesthesia led to lower intra-
were more likely to be given spinal anaesthesia. There
operative blood pressure and therefore a reduction in
may, therefore, be support within the current evidence for
blood loss [17]. However, the findings of the anaesthetic
an association between spinal anaesthesia and an
sprint audit of practice suggested that intra-operative
increased risk of thromboembolic events in hip fracture
arterial blood pressure was lower with general anaesthesia
operations.
in patients having hip fracture surgery [6]. In addition, our
The increase in postoperative chest infections within
results are concordant with a meta-analysis of 66
the spinal anaesthetic group is probably a consequence of
randomised controlled trials that found the mean
two opposing factors. Spinal anaesthesia is generally
difference of estimated blood loss to be 335 ml lower with
believed to be protective or neutral for respiratory
spinal compared with general anaesthesia in a variety of
complications [14, 15] but this in turn may lead to greater
surgical procedures [18]. It is also possible that a degree of
use of spinal anaesthesia in patients at risk of respiratory
residual confounding may have led to this finding as no
complications [24]. The proportion of patients with COPD
data were available to match patients based on existing
who received spinal anaesthesia was almost double that of
coagulopathy; this condition is a relative contra-indication
those who had a general anaesthetic; this is likely to reflect
to spinal anaesthesia, due to the increased risk of vertebral
deliberate clinical decision-making to avoid general
canal haematoma and subsequent spinal cord compression
anaesthesia in patients with pre-existing respiratory
[19]. Given that patients with an existing coagulopathy
disease. However, this introduces the possibility of
have an increased likelihood of requiring a postoperative
baseline confounding; those patients with existing COPD
transfusion [20], and are also more likely to be deemed
are more likely to develop a chest infection
unsuitable for a spinal anaesthetic, it is possible that this
postoperatively and are relatively overrepresented in the
had a confounding effect.
population who have a spinal anaesthetic. A clinical
The association of spinal anaesthetic with a reduction
argument is made frequently that patients with specific
in
postoperative UTIs was surprising. There is very little underlying pathology, particularly COPD, are better served

evidence in the literature that directly supports or by spinal anaesthesia. We could not find any convincing

contradicts this finding, as previous studies have tended to evidence of this benefit. As with all non-randomised

focus on postoperative urinary retention (which may studies, we cannot exclude residual confounding,

subsequently require catheterisation and increase the risk particularly regarding severity of COPD.

of the patient developing a UTI). Some of these studies There were limitations to this study which should be

have suggested that spinal anaesthesia increases urinary acknowledged. The pre-operative clinical characteristics

retention compared with general anaesthesia. However, and postoperative measures of morbidity recorded were

most urinary catheters in our unit are placed pre- not always clearly defined. As such, there is the potential

operatively in accordance with departmental protocols. for misclassification bias if data were recorded in different

The research in this area is inconsistent [21, 22] and the ways as a result of differences in interpretation and/or

true relative effects of spinal and general anaesthesia on diagnosis between individual clinicians. The timeframe

urinary retention and UTIs are unclear. It is possible that over which data were recorded into the database and the

catheterisation rates may have altered the clinical changes in clinical practice during said time-period would

recording of a diagnosis of UTI, which is notoriously itself have potentially affected the concordance of

unreliable. Positive dipstick tests in a patient who is different clinicians’ opinions, let alone the inherent

catheterised may be ascribed less importance than if the variation that would exist amongst practitioners

patient was not catheterised. regardless. In addition, despite the

Morgan et al. | Effects of anaesthetic technique in hip fracture surgery Anaesthesia 2020
© 2020 The Authors. Anaesthesia published by John Wiley & Sons Ltd on behalf of Association of Anaesthetists 5
Anaesthesia 2020 Morgan et al. | Effects of anaesthetic technique in hip fracture surgery
advantages of the propensity score-matching technique as
a method of adjusting to limit the effects of confounding, it
is possible that it may have actually increased the risk of
confounding, as a result of an effect described as ‘the
propensity score paradox’ [25, 26]. Within the context of
this study, in which the covariates were already well-
balanced between the two groups, the ‘paradox’ describes
how the technique of propensity score-matching may have
led to imbalance within the distribution of the clinical
characteristics; this occurs through the pruning of pairs
with the largest difference in propensity scores, and
potentially introduces bias into the analysis [26]. In
addition, the dataset does not include information about
delirium rates or changes in cognitive function. Although
the data are of limited quality, this may be where a benefit
of spinal anaesthesia lies (if there is one to be found).
Another limitation is that spinal and general anaesthesia
have been considered as single entities; we were unable to
consider how these were delivered. The study may have
been comparing ‘bad’ spinal anaesthesia with ‘good’
general anaesthesia or vice versa. Similarly, we are unable
to comment on the impact of sedation used in conjunction
with spinal anaesthesia [27, 28]. Until routine data
collection allows analysis of how anaesthesia was
delivered, these questions are likely to remain
unanswered. Finally, but perhaps most importantly, we
may be simply measuring the wrong outcomes [29]. If
clinical outcomes are unaffected by mode of anaesthesia,
then addressing softer outcomes such as patient
satisfaction, quality of life and operating theatre efficiency
may be more important.
Despite these limitations, there is a major strength: with
the power of an 8000-patient dataset, and data on a
multitude of outcomes, we were unable to demonstrate
clinically- relevant differences between modes of anaesthesia.
For individual clinicians to state that his or her favoured
technique is optimal based on incomplete data involving only
tens or hundreds of patients is an unprovable claim.
There are four ongoing randomised controlled trials of
spinal vs. general anaesthesia [30–33]. We await the
results of these with interest. In the meantime, it would
seem the decision to use a particular mode of anaesthesia
for a patient undergoing hip fracture surgery is primarily
based around patient choice and the inclinations of
individual practitioners [34], rather than a standardised
approach. We suggest that future research should be
directed towards how anaesthesia is delivered, rather than
20. Liumbruno GM, Bennardello F, Lattanzio A, Piccoli P, Rossetti
mode of anaesthesia per se.
Acknowledgements
No external funding or competing interests declared.
References
1. Petrisor B, Bhandari M. BMJ best practice: hip fractures. 2018.
https://bestpractice.bmj.com/topics/en-gb/387 (accessed 27/
09/2018).
2. Royal College of Physicians. National Hip Fracture Database
annual report. 2017. https://www.nhfd.co.uk/files/2017Re
portFiles/NHFD-AnnualReport2017.pdf (accessed 27/09/
2018).
3. White S, Griffiths R. Projected incidence of proximal femoral fracture
in England: a report from the NHS Hip Fracture Anaesthesia
Network. Injury 2011; 42: 1230–3.
4. Kanonidou Z, Karystianou G. Anesthesia for the elderly.
Hippokratia 2007; 11: 175–7.
5. Bowker L, Price J, Smith S. Oxford handbook of geriatric medicine,
2nd edn. Oxford, UK: Oxford University Press, 2012.
6. Association of Anaesthetists of Great Britain and Ireland.
Anaesthetic sprint audit of practice. 2014. https://www.nhfd.co.
uk/20/hipfractureR.nsf/vwContent/asapReport (accessed 20/
02/2020).
7. Bodenham A, Howell S. General anaesthesia vs local anaesthesia:
an ongoing story. British Journal of Anaesthesia 2009; 103: 785–9.
8. Royal College of Anaesthetists. Your spinal anaesthetic:
information for patients. 2014. https://www.rcoa.ac.uk/sites/
default/files/documents/2020-01/03-YourSpinalweb.pdf (accessed
28/02/2020).
9. National Institute for Health and Care Excellence. Hip fracture:
management. 2011.
https://www.nice.org.uk/guidance/cg124/ resources/hip-
fracture-management-pdf-35109449902789 (accessed
28/02/2020).
10. Maxwell MJ, Moran CG, Moppett IK. Development and
validation of a preoperative scoring system to predict 30-day
mortality in patients undergoing hip fracture surgery. British
Journal of Anaesthesia 2008; 101: 511–7.
11. Buss L, McKeever TM, Nightingale J, et al. Hip fracture
outcomes in patients with chronic obstructive pulmonary disease.
British Journal of Anaesthesia 2018; 121: 1377–9.
12. Sagar J, Blott H, Nisar P, et al. Abbreviated Mental Test
(cognitive) screening in acutely ill elderly surgical patients.
British Journal of Surgery 2015; 102: 137–8.
13. Thavaneswaran A, Lix L. Propensity score matching in
observational Studies: University of Manitoba. 2008. https://
umanitoba.ca/faculties/health_sciences/medicine/units/chs/
departmental_units/mchp/protocol/media/propensity_sc
ore_matching.pdf (accessed 28/02/2020).
14. Guay J, Parker MJ, Gajendragadkar PR, Kopp S. Anaesthesia for
hip fracture surgery in adults. Cochrane Database of Systematic
Reviews 2016; 2: 1465–858.
15. Van Waesberghe J, Stevanovic A, Rossaint R, Coburn M.
General vs. neuraxial anaesthesia in hip fracture patients: a
systematic review and meta-analysis. BMC Anesthesiology
2017; 17: 1–25.
16. Neuman MD, Rosenbaum PR, Ludwig JM, et al. Anesthesia
technique, mortality, and length of stay after hip fracture surgery.
Journal of American Medical Association 2014; 311: 2508–17.
17. Rashiq S, Finegan BA. The effect of spinal anesthesia on blood
transfusion rate in total joint arthroplasty. Canadian Journal of
Surgery 2006; 49: 391–6.
18. Richman JM, Rowlingson AJ, Maine DN, et al. Does neuraxial
anesthesia reduce intraoperative blood loss? A meta-analysis.
Journal of Clinical Anesthesia 2006; 18: 427–35.
19. Harrop-Griffiths W, Cook T, Gill H, et al. Regional anaesthesia
and patients with abnormalities of coagulation. Anaesthesia
2013; 68: 966–72.
G. Recommendations for the transfusion management of
6 © 2020 The Authors.
Anaesthesia published by John Wiley & Sons Ltd on behalf of Association of
Anaesthetists
Morgan et al. | Effects of anaesthetic technique in hip fracture surgery Anaesthesia 2020

patients in the peri-operative period-the postoperative period.

28. Sieber FE, Neufeld KJ, Gottschalk A, et al. Effect of depth of
Blood Transfusion 2011; 9: 320–35.
sedation in older patients undergoing hip fracture repair on
21. Chu CPW, Yap JCCM, Chen PP, Hung HH. Postoperative
postoperative delirium: the STRIDE randomized clinical trial.
outcome in Chinese patients having primary total knee
Journal of the American Medical Association Surgery 2018; 153:
arthroplasty under general anaesthesia/intravenous patient-
987–95.
controlled analgesia compared to spinal-epidural anaesthesia/
29. Moppett I. Hip fractures: are we asking the right questions?
analgesia. Hong Kong Medical Journal 2006; 12: 442.
Age and Ageing 2018; 47: 633–4.
22. Lingaraj K, Ruben M, Chan YH, De SD. Identification of risk
30. ISRCTN Registry. REGARD: comparing regional and general
factors for urinary retention following total knee arthroplasty:
anaesthesia and their effect on delirium in patients with hip
a Singapore hospital experience. Singapore Medical Journal
fractures. 2017. http://www.isrctn.com/ISRCTN15165914
2007; 48: 213–6.
(accessed 28/02/2020).
23. National Institute for Health and Care Excellence. Venous
31. Neuman MD, Ellenberg SS, Sieber FE, et al. Regional versus
thromboembolism in over 16s: reducing the risk of hospital-
General Anesthesia for Promoting Independence after Hip
acquired deep vein thrombosis or pulmonary embolism. 2018.
Fracture (REGAIN): protocol for a pragmatic, international
https://www.nice.org.uk/guidance/ng89 (accessed 28/02/2020).
multicentre trial. British Medical Journal Open 2016; 6:
24. Conde M, Lawrence V. Postoperative pulmonary infections.
e013473.
BMJ Clinical Evidence 2008; 2008: 2201.
32. Kowark A, Adam C, Ahrens J, et al. Improve hip fracture
25. King G, Nielsen R. Why propensity scores should not be used
outcome in the elderly patient (iHOPE): a study protocol for a
for matching. 2018. https://gking.harvard.edu/files/gking/file
pragmatic, multicentre randomised controlled trial to test the
s/pan1900011_rev.pdf (accessed 28/02/2020).
efficacy of spinal versus general anaesthesia. British Medical
26. Ripollone JE, Huybrechts KF, Rothman KJ, Ferguson RE,
Journal Open 2018; 8: e023609.
Franklin JM. Implications of the propensity score matching
33. Li T, Yeung J, Li J, et al. Comparison of regional with general
paradox in pharmacoepidemiology. American Journal of
anaesthesia on postoperative delirium (RAGA-delirium) in the
Epidemiology 2018; 187: 1951.
older patients undergoing hip fracture surgery: study protocol
27. Sieber FE, Zakriya KJ, Gottschalk A, et al. Sedation depth
for a multicentre randomised controlled trial. British Medical
during spinal anesthesia and the development of
Journal Open 2017; 7: e016937.
postoperative delirium in elderly patients undergoing hip
34. White SM, Griffiths R, Moppett IK. Standardising anaesthesia
fracture repair. Mayo Clinic Proceedings 2010; 85: 18–26.
for hip fracture surgery. Anaesthesia 2016; 71: 1391–5.

© 2020 The Authors. Anaesthesia published by John Wiley & Sons Ltd on behalf of Association of Anaesthetists 7