Background:

Trigeminal neuralgia (TN) is a well-recognized cause of facial pain in the general population, and multiple sclerosis (MS) accounts for some of these cases. However, the prevalence of TN in MS is poorly understood. We investigated the prevalence of TN and how often TN is the initial presentation of MS in a large MS cohort.

Methods:

In 2009, we surveyed participants in the North America Research Committee on Multiple Sclerosis Registry regarding TN, including date of onset and pharmacologic and nonpharmacologic treatments used. We estimated the frequency of TN and the frequency with which TN preceded the diagnosis of MS. We compared the demographic and clinical characteristics of participants who reported TN with those of participants who did not using descriptive statistics and logistic regression.

Results:

Among 8590 eligible survey respondents, the prevalence of TN was 830 (9.7%). Of these respondents, 588 reported the year when TN was diagnosed. The diagnosis of TN preceded that of MS in 88 respondents (15.0%), and the mean ± SD age at diagnosis of TN was 45.3 ± 11.0 years. The odds of reporting TN were higher in women and those with greater disability and longer disease duration. Pharmacologic treatments were used by 88.3% of respondents; 9.7% underwent surgical interventions.

Conclusions:

In MS, TN occurs frequently and precedes the diagnosis of MS in 15.0% of individuals. Given the frequency of TN in MS, further epidemiological studies and clinical trials to identify effective pharmacologic and nonpharmacologic therapies for TN in MS are warranted.

Trigeminal neuralgia (TN) is a well-recognized cause of facial pain in the general population, although it is a relatively rare condition with a lifetime prevalence of up to 0.3%.1 Trigeminal neuralgia is defined as a recurrent unilateral brief electric shock–like pain, abrupt in onset and termination, limited to the distribution of one or more divisions of the trigeminal nerve and triggered by innocuous stimuli.2 The pain is usually triggered by sensory stimuli to the skin, mucosa, or teeth that are innervated by the ipsilateral trigeminal nerve. Although each attack lasts for a few seconds, the pain can be repetitive at short intervals so that the individual attacks blur into one another.

The prevalence of pain in MS is high, affecting more than 60% of individuals during the course of their disease.3 A recent meta-analysis reported that the prevalence of TN in MS was 3.8% (95% confidence interval [CI], 2%–6%) based on seven studies with heterogeneous findings. Although the occurrence of TN secondary to MS is now referred to as painful TN attributed to MS plaque,2 it will be referred to as TN throughout this article for simplicity. The prevalence of specific pain syndromes, such as TN, before and at MS onset was noted to be particularly poorly described in the literature.3 A single-center study conducted in British Columbia, Canada, reported that only five of 1182 individuals evaluated from 1980 to 1992 had TN as the first presentation of MS.4 

Given the limited understanding of the prevalence of TN in MS, we aimed to investigate its prevalence in a large MS population, hypothesizing that the prevalence of TN would be higher than previously reported. We specifically aimed to investigate the prevalence of TN presenting before the diagnosis of MS.

Methods

Study Population

We conducted this study using data from participants in the North American Research Committee on Multiple Sclerosis (NARCOMS) Registry. Since 1996, 37,804 individuals have enrolled in the NARCOMS Registry. At enrollment, participants complete a questionnaire online or on paper in which they report demographic (including gender, date of birth, race, education, and annual household income) and clinical information (including year of MS symptom onset, year of diagnosis, type of MS, disability status); they update clinical information semi-annually by questionnaire. Diagnoses of MS have been validated in a randomly selected sample, as have participant-reported assessments of their type of MS.5 Disability status was reported using the Patient-Determined Disease Steps (PDDS) scale, a validated measure that is highly correlated with the Expanded Disability Status Scale score as recorded after physician assessment.6,7 The PDDS scale is scored from 0 (no disability) to 8 (bedridden). Fatigue, cognition, spasticity, sensory, bladder, vision, and hand function were assessed using Performance Scales (DeltaQuest Foundation Inc., Concord, MA).8 Each of these domains is scored as follows: 0 (normal), 1 (minimal), 2 (mild), 3 (moderate), 4 (severe), or 5 (total disability).

The institutional review board of the University of Alabama at Birmingham approves the NARCOMS Registry.

Trigeminal Neuralgia

In the fall of 2009, the semiannual survey asked participants if they have ever been diagnosed with TN, the year of TN diagnosis, and the treatments used to manage TN in the past and in the previous 6 months. To improve the accuracy of reporting, participants were given the following description of TN: “Trigeminal neuralgia (TN), also known as tic douloureux (TD), is a neuropathic disorder characterized by an intermittent sharp and severe pain affecting generally only one side of the face and generally occurring in ‘spells’ lasting days to weeks.” Participants were not asked whether they had any other new neurologic symptoms at the time TN was diagnosed.

Regarding drug therapies, participants reported the use of anticonvulsants (carbamazepine, gabapentin, oxcarbazepine, phenytoin, lamotrigine, pregabalin, valproate, topiramate, clonazepam), tricyclic antidepressants (amitriptyline, nortriptyline), antispasmodics (tizanidine, baclofen), opioids (morphine, oxycodone, hydrocodone, hydromorphone, codeine, tramadol), and topical lidocaine. They also reported the use of interventions administered by a health professional, including botulinum toxin, microvascular decompression, gamma-knife radiosurgery, thermocoagulation, mechanical balloon compression, and glycerol injections.

Analysis

For this analysis, we excluded individuals who did not answer the question regarding whether they had a diagnosis of TN or who indicated a response of unsure. To limit heterogeneity due to potential differences in adult-onset and pediatric-onset MS, we excluded those with an age at MS symptom onset of 16 years or younger. Analyses of the timing of TN diagnosis were limited to participants with TN who reported the year of diagnosis.

We summarized the demographic and clinical characteristics of participants using means, standard deviations (SDs), medians, interquartile ranges, frequencies, and percentages, as appropriate. We report the proportion of participants with a diagnosis of TN and when this diagnosis occurred in relation to the diagnosis of MS. We also report the frequency with which specific treatments for TN were used. Student t, Wilcoxon, and χ2 tests were used to compare the demographic characteristics of participants with and without a diagnosis of TN and those of participants with a diagnosis of TN before or after MS diagnosis.

We then conducted multivariable logistic regression to compare the characteristics of participants with and without a diagnosis of TN. Covariates considered were age, disease duration, sex, race, annual household income, disability, and functional status. Age and disease duration were categorized into quintiles, with the lowest quintile as the reference group. Race was included as white versus nonwhite (reference group). Annual household income was categorized as $15,000 or less, $15,001 to $30,000, $30,001 to $50,000, $50,001 to $100,000, greater than $100,000 (reference group), and prefer not to answer. Disability status was categorized as mild (PDDS score = 0–2 [reference group]), moderate (PDDS score = 3), or severe (PDDS score = 4–8).9 Functional domains were categorized as minimal (0–1) (reference group), mild-moderate (2–3), and severe (4–5).

Statistical analyses were performed using IBM SPSS Statistics for Windows, version 23.0 (IBM Corp., Armonk, NY) and SAS, version 9.3 (SAS Institute Inc., Cary, NC).

Results

The fall 2009 survey was distributed to 13,142 participants, of whom 10,102 responded (76.9%). Responders and nonresponders did not differ with respect to sex, age at MS symptom onset, or level of disability. Responders were slightly older (mean ± SD age 54.3 ± 10.3 vs. 53.4 ± 11.5 years, P < .0001) than nonresponders.

Of those who responded, 8590 were included in the analysis (Figure 1). Most participants were female and white, with a mean ± SD age at MS onset of 31.7 ± 9.1 years (Table 1). Nearly 10% of participants (830 of 8590) reported a diagnosis of TN. On univariate analysis, compared with participants without TN, those with TN were more likely to be female, to have an earlier age at symptom onset and a longer disease duration, and to experience greater disability in all the domains assessed (Table 1). Race, annual income, and age at MS diagnosis did not differ between groups.

Figure 1.

Flow diagram showing selection of study population

MS, multiple sclerosis; TN, trigeminal neuralgia.

Figure 1.

Flow diagram showing selection of study population

MS, multiple sclerosis; TN, trigeminal neuralgia.

Table 1.

Characteristics of study population

Characteristics of study population
Characteristics of study population

On multivariable logistic regression analysis, several characteristics were independently associated with increased odds of reporting a diagnosis of TN, including longer disease duration and greater impairment in the domains of hand function, vision, cognition, and fatigue (Table 2). Male sex was associated with decreased odds of reporting TN.

Table 2.

Odd ratios for associations between participant characteristics and diagnosis of trigeminal neuralgia

Odd ratios for associations between participant characteristics and diagnosis of trigeminal neuralgia
Odd ratios for associations between participant characteristics and diagnosis of trigeminal neuralgia

In the 588 participants with TN who reported the year that TN was diagnosed, TN preceded the diagnosis of MS in 88 (15.0%), and occurred the same year as the diagnosis of MS in 104 (17.7%). On univariate analysis, the two groups did not differ with respect to sex, race, income, or disability (PDDS scale score) (Table 3). Age at MS diagnosis was an average of 4.3 years later in those with TN diagnosed before MS than in those with TN first diagnosed after MS was diagnosed (P = .0001).

Table 3.

Characteristics of participants with TN diagnosed before versus after MS diagnosis

Characteristics of participants with TN diagnosed before versus after MS diagnosis
Characteristics of participants with TN diagnosed before versus after MS diagnosis

As expected, use of pharmacologic treatments for TN was common, with 733 respondents (88.3%) reporting that they had ever used one of the therapies queried. The class of medications that was most commonly ever used was anticonvulsants (655 [78.9%]), followed by antispasmodics (422 [50.8%]) and opioids (306 [36.9%]). The most commonly used medication was gabapentin, followed by carbamazepine (Table 4). Nearly 65% of respondents had used one or more pharmacologic therapies in the past 6 months (n = 537). Of these, 230 (42.8%) used one, 147 (27.4%) used two, 93 (17.3%) used three, and 67 (12.5%) used four or more of these pharmacologic therapies. With respect to procedures, 81 respondents (9.7%) had ever undergone at least one procedure to treat their TN. The most common procedure reported was gamma-knife radiotherapy, followed by glycerol injections (Table 5).

Table 4.

Pharmacologic treatments for trigeminal neuralgia

Pharmacologic treatments for trigeminal neuralgia
Pharmacologic treatments for trigeminal neuralgia
Table 5.

Surgical interventions for trigeminal neuralgia

Surgical interventions for trigeminal neuralgia
Surgical interventions for trigeminal neuralgia

Discussion

Pain is a common symptomatic concern in MS; a recent meta-analysis of 28 studies reported a pooled prevalence of 63% (95% CI, 55%–70%).3 Multiple pain syndromes have been described, including painful spasms, headache, Lhermitte sign, and TN,3 but the last is understudied. In this large study, we found that TN was common in people with MS, affecting 9.5% of individuals over the course of their disease. Furthermore, TN preceded the diagnosis of MS in 15% of those individuals and occurred in the same year as the diagnosis of MS in another 17.7%. Individuals with a history of TN were more likely to be female, to have a longer disease duration, and to report greater impairments regarding hand function, vision, cognition, and fatigue than those without a history of TN. Most participants with TN used pharmacologic therapy to manage it.

The prevalence of TN in this sample is higher than that in previous reports. In the previously mentioned meta-analysis of the prevalence of pain in MS, the pooled prevalence of TN in seven studies was 3.8% (95% CI, 2%–6%).3 Sample sizes in these studies ranged from 28 to 1672, and some of the samples were drawn from hospitalized populations or single centers, limiting generalizability. The largest of these studies included participants from multiple centers but reported a low prevalence of TN of only 2%.10 This may reflect the focus on current pain symptoms at the time the study was conducted rather than on lifetime diagnoses of TN, as reported in the present study. In one study in which both current and lifetime TN were reported, the lifetime prevalence of TN was 4.0%,11 four-fold higher than the current prevalence (1.2%). Our focus on lifetime prevalence in individuals with an average duration of disease of more than 20 years may have enabled us to capture a more accurate lifetime estimate.

The diagnosis of TN preceded the diagnosis of MS in 15% of the study population, similar to what has been reported previously.4 The literature regarding the prevalence of TN relative to the timing of MS diagnosis is limited.3 Hooge and Redekop4 evaluated the medical records of 1882 patients from 1980 until 1992 and found that only five (14.3% of those with TN) had a diagnosis of TN before MS diagnosis. The mean age at diagnosis of TN was 45.3 years in the present cohort, similar to previous reports by Rushton and Olafson12 (45.2 years) and Jensen et al.13 (43.4 years) but lower than that reported by Hooge and Redekop4 (51 years).4 Most studies that have compared the age at diagnosis of TN in individuals with and without MS have found that TN occurs at younger ages in those with MS.13,14 

Several factors were associated with a greater likelihood of reporting TN. As expected, longer disease duration was associated with greater odds of reporting TN. Men were less likely to report TN than were women, even after accounting for age, disease duration, and other factors. Although women with MS have a higher risk of headache disorders,11 previous studies of factors associated with prevalent pain disorders generally either have not evaluated the effects of sex15 or have not shown any association with sex.10 In studies of TN, the association with sex has not been evaluated or has not been identified, but the number of patients with TN has been small.4,12,16 The associations with greater reported impairment in the domains of hand function, vision, cognition, and fatigue are more difficult to interpret and, to our knowledge, have not been described previously. Owing to the cross-sectional design of this study, we cannot determine whether these perceived impairments reflect adverse effects of medications used for TN (such as impaired cognition and greater fatigue) or the impact of pain due to TN (fatigue) or other factors, such as a greater propensity for lesions affecting domains such as vision. These issues warrant replication and further study as to the underlying mechanisms.

We found that 88.3% of participants had ever used pharmacologic therapy for TN, and 78% of these individuals continued to use pharmacologic therapy in the past 6 months. Use of multiple pharmacologic therapies was common, with more than half of those who reported treatment in the past 6 months using two or more agents. The most common pharmacologic agent ever used was gabapentin, followed by carbamazepine. Use of opioids was also fairly common. In classical TN, carbamazepine is considered an effective, first-line treatment.17 Oxcarbazepine is considered probably effective, and baclofen, lamotrigine, and pimozide are considered possibly effective. In MS, randomized, double-blind, placebo-controlled clinical trials of pharmacologic therapies for TN are lacking. Controlled clinical trials have suggested that carbamazepine is helpful for paroxysmal pain syndromes, including TN.18,19 Other studies, generally small open-label trials and observational studies, suggest that gabapentin reduces pain20,21 and that misoprostol may also be effective.22 Thus, many of the therapies tried by the present study population have not been shown to be effective for TN. The need for multiple agents suggests that these therapies are not completely effective in alleviating TN.

Nearly 10% of respondents had undergone at least one procedure for TN, most often gamma-knife radiotherapy. In individuals with TN refractory to pharmacologic therapy, surgical interventions are often considered, including microvascular decompression,23 gamma-knife radiosurgery,24 and percutaneous balloon compression,25 although clinical trial evidence to support these interventions is lacking in MS and case series suggest that the effectiveness and risk of adverse effects may differ from those seen in the classical TN population.24,25 This highlights the need for additional clinical trials of treatments for TN.

This study has several limitations. Participants in the NARCOMS Registry are volunteers, potentially reducing generalizability of the findings. The response rate was 76.9%, and nonresponders differed from responders in age and disability level, although these differences were so small as to have limited clinical meaning. Although participants in the NARCOMS Registry have been shown to accurately report their diagnoses of MS, years of symptom onset and diagnosis, clinical course, and comorbidities,5 we did not validate the accuracy of the question used to ascertain a reported diagnosis of TN. Participants may have mistakenly reported facial pain or other headache disorders that were not TN as TN, but we provided a description of TN before asking questions about this pain syndrome, and other studies that have evaluated the prevalence of TN in MS have also used questionnaires.10 We did not ask participants to indicate whether TN presented as an isolated (monofocal) symptom or as part of a polyfocal presentation, and this should be pursued in future studies. Nonetheless, the pattern of use of pharmacologic therapy is consistent with expectations for TN. A large number of participants did not report the year of TN diagnosis, introducing some uncertainty in the estimated prevalence of TN before MS diagnosis. For example, if we make the extreme assumption that TN occurred after MS diagnosis in all participants in whom the year of TN diagnosis was not reported, then TN preceded the diagnosis of MS in only 10.6% of the participants (88 of 830), but this still makes TN a common presenting symptom of MS. The study design was not appropriate to evaluate participants' perceptions of the effectiveness of these therapies.

The present findings indicate that TN is relatively common in the MS population and that the risk increases with disease duration. In 15% of the MS population, TN precedes the diagnosis of MS. A wide range of therapies are used to treat TN, but many of these therapies have not been adequately tested in clinical trials. Given the frequency of TN in MS, further trials to test pharmacologic and nonpharmacologic therapies for TN in MS are warranted.

PracticePoints

  • Trigeminal neuralgia affects 15% of individuals before the diagnosis of MS.

  • In MS, the prevalence of trigeminal neuralgia increases with disease duration and disability.

Financial Disclosures

Dr. Cutter has served on data and safety monitoring boards for AMO Pharma, Apotek, Biogen-Idec, Glaxo-SmithKline Pharmaceuticals, Gilead Pharmaceuticals, Horizon Pharmaceuticals, ModigeneTech/Prolor, Merck Pharmaceuticals, Opko, Neuren, Sanofi-Aventis, Teva, the National Heart, Lung, and Blood Institute (protocol review committee), and the National Institute of Child Health and Human Development (Obstetric-Fetal Pharmaceutical Research Units oversight committee); and has received consulting fees from and/or served on scientific advisory boards for CereSpir Inc., the Consortium of Multiple Sclerosis Centers (CMSC), D3, Genzyme, Genentech, Innate Therapeutics, Janssen Pharmaceuticals, Klein Buendel Inc, MedImmune, MedDay, Novartis, Opexa Therapeutics, Receptos, Roche, Savara Inc., Spinifex Pharmaceuticals, Somahlution, Teva Pharmaceuticals, Transparency Life Sciences, and TG Therapeutics. Dr. Marrie receives research funding from the Canadian Institutes of Health Research, Research Manitoba, Multiple Sclerosis Society of Canada, Multiple Sclerosis Scientific Foundation, National Multiple Sclerosis Society, Rx & D Health Research Foundation, and the Waugh Family Chair in Multiple Sclerosis and has conducted clinical trials funded by Sanofi-Aventis. The other authors have no conflicts of interest to disclose.

Funding/Support

The NARCOMS Registry is supported, in part, by the CMSC and the Foundation of the CMSC. Dr. Marrie was supported, in part, by a Don Paty Career Development Award from the MS Society of Canada, a Manitoba Research Chair from Research Manitoba, and the Waugh Family Chair in Multiple Sclerosis. Performance Scales, Copyright Registration Number / Date: TXu000743629 / 1996-04-04; assigned to DeltaQuest Foundation, Inc., effective October 1, 2005. U.S. Copyright law governs terms of use.

References

References
1.
Mueller
D
,
Obermann
M
,
Yoon
M-S
,
et al
.
Prevalence of trigeminal neuralgia and persistent idiopathic facial pain: a population-based study
.
Cephalalgia.
2011
;
31
:
1542
1548
.
2.
Headache Classification Committee of the International Headache Society (IHS)
.
The International Classification of Headache Disorders, 3rd ed (beta version)
.
Cephalalgia.
2013
;
33
:
629
808
.
3.
Foley
PL
,
Vesterinen
HM
,
Laird
BJ
,
et al
.
Prevalence and natural history of pain in adults with multiple sclerosis: systematic review and meta-analysis
.
Pain.
2013
;
154
:
632
642
.
4.
Hooge
JP
,
Redekop
WK.
Trigeminal neuralgia in multiple sclerosis
.
Neurology.
1995
;
45
:
1294
1296
.
5.
Marrie
RA
,
Cutter
G
,
Tyry
T
,
Campagnolo
D
,
Vollmer
T.
Validation of the NARCOMS Registry: diagnosis
.
Mult Scler.
2007
;
13
:
770
775
.
6.
Marrie
RA
,
Goldman
MD.
Validity of Performance Scales for disability assessment in multiple sclerosis
.
Mult Scler.
2007
;
13
:
1176
1182
.
7.
Learmonth
YC
,
Motl
RW
,
Sandroff
BM
,
Pula
JH
,
Cadavid
D.
Validation of Patient Determined Disease Steps (PDDS) scale scores in persons with multiple sclerosis
.
BMC Neurol.
2013
;
13
:
37
.
8.
Schwartz
CE
,
Vollmer
T
,
Lee
H
,
North American Research Consortium on Multiple Sclerosis Outcomes Study Group
.
Reliability and validity of two self-report measures of impairment and disability for MS
.
Neurology.
1999
;
52
:
63
71
.
9.
Marrie
RA
,
Cutter
G
,
Tyry
T
,
Vollmer
T
,
Campagnolo
D.
Does multiple sclerosis-associated disability differ between races?
Neurology.
2006
;
66
:
1235
1240
.
10.
Solaro
C
,
Brichetto
G
,
Amato
MP
,
et al
.
The prevalence of pain in multiple sclerosis: a multicenter cross-sectional study
.
Neurology.
2004
;
63
:
919
921
.
11.
Martinelli Boneschi
F
,
Colombo
B
,
Annovazzi
P
,
et al
.
Lifetime and actual prevalence of pain and headache in multiple sclerosis
.
Mult Scler.
2008
;
14
:
514
521
.
12.
Rushton
JG
,
Olafson
RA.
Trigeminal neuralgia associated with multiple sclerosis: report of 35 cases
.
Arch Neurol.
1965
;
13
:
383
386
.
13.
Jensen
TS
,
Rasmussen
P
,
Reske-Nielsen
E.
Association of trigeminal neuralgia with multiple sclerosis: clinical and pathological features
.
Acta Neurol Scand.
1982
;
65
:
182
189
.
14.
De Simone
R
,
Marano
E
,
Brescia Morra
V
,
et al
.
A clinical comparison of trigeminal neuralgic pain in patients with and without underlying multiple sclerosis
.
Neurol Sci.
2005
;
26
(
suppl 2
):
S150
S151
.
15.
Vermote
R
,
Ketelaer
P
,
Carton
H.
Pain in multiple sclerosis patients: a prospective study using the McGill Pain Questionnaire
.
Clin Neurol Neurosurg.
1986
;
88
:
87
93
.
16.
Österberg
A
,
Boivie
J
,
Thuomas
KÅ.
Central pain in multiple sclerosis: prevalence and clinical characteristics
.
Eur J Pain.
2005
;
9
:
531
542
.
17.
Gronseth
G
,
Cruccu
G
,
Alksne
J
,
et al
.
Practice parameter: the diagnostic evaluation and treatment of trigeminal neuralgia (an evidence-based review): report of the Quality Standards Subcommittee of the American Academy of Neurology and the European Federation of Neurological Societies
.
Neurology.
2008
;
71
:
1183
1190
.
18.
Rockliff
BW
,
Davis
EH.
Controlled sequential trials of carbamazepine in trigeminal neuralgia
.
Arch Neurol.
1966
;
15
:
129
136
.
19.
Espir
ML
,
Millac
P.
Treatment of paroxysmal disorders in multiple sclerosis with carbamazepine (Tegretol)
.
J Neurol Neurosurg Psychiatry.
1970
;
33
:
528
531
.
20.
Khan
OA.
Gabapentin relieves trigeminal neuralgia in multiple sclerosis patients
.
Neurology.
1998
;
51
:
611
614
.
21.
Solaro
C
,
Messmer Uccelli
M
,
Uccelli
A
,
Leandri
M
,
Mancardi
GL.
Low-dose gabapentin combined with either lamotrigine or carbamazepine can be useful therapies for trigeminal neuralgia in multiple sclerosis
.
Eur Neurol.
2000
;
44
:
45
48
.
22.
DMKG Study Group
Misoprostol in the treatment of trigeminal neuralgia associated with multiple sclerosis
.
J Neurol.
2003
;
250
:
542
545
.
23.
Broggi
G
,
Ferroli
P
,
Franzini
A
,
Pluderi
M
,
La Mantia
L
,
Milanese
C.
Role of microvascular decompression in trigeminal neuralgia and multiple sclerosis
.
Lancet.
1999
;
354
:
1878
1879
.
24.
Huang
E
,
Teh
BS
,
Zeck
O
,
et al
.
Gamma knife radiosurgery for treatment of trigeminal neuralgia in multiple sclerosis patients
.
Stereotact Funct Neurosurg.
2002
;
79
:
44
50
.
25.
Martin
S
,
Teo
M
,
Suttner
N.
The effectiveness of percutaneous balloon compression in the treatment of trigeminal neuralgia in patients with multiple sclerosis
.
J Neurosurg.
2015
;
123
:
1507
1511
.

Author notes

From the Department of Internal Medicine (AF, RAM) and the Department of Community Health Sciences (RAM), University of Manitoba, Winnipeg, Manitoba, Canada; Department of Biostatistics, Washington University in St. Louis, St. Louis, MO, USA (AS); Dignity Health, St. Joseph's Hospital and Medical Center, Phoenix, AZ, USA (TT); University of Alabama at Birmingham, Birmingham, USA (GRC).