1

Early clinical sequelae of Ebola virus disease in Sierra Leone: a cross-sectional study 1

2

John G. Mattia, MBBS1,2; Mathew J. Vandy, MBBS1; Joyce C. Chang, RN3; Devin E. Platt, 3

BSN3; Kerry Dierberg, MD3, Daniel G. Bausch, MD4; Tim Brooks, MBChir5; Sampha Conteh, 4

NDSW3;Ian Crozier, MD6,7; Robert A. Fowler, MDCM6,8; Amadu P. Kamara, CON1,2; Cindy 5

Kang, PA3; Srividya Mahadevan, MPH9; Yealie Mansaray,CON1,2; Lauren Marcell,BA3; Gillian 6

McKay, MScPH9; Tim O’Dempsey, MRCP6,10; Victoria Parris, MRCP11; Ruxandra Pinto, 7

PhD12; Audrey Rangel, MPH11; Alex P. Salam, MBChB9; Jessica Shantha, MD13; Vanessa 8

Wolfman, MD11; Steven Yeh, MD13; Adrienne K. Chan, MD6,8,15; Sharmistha Mishra, MD6,8 9

10

1National Eye Health Program, Ministry of Health and Sanitation, Government of Sierra Leone, 11Freetown, Sierra Leone 122Christian Blind Mission, Baptist Eye Hospital Lunsar and Lowell and Ruth Gess UMC Eye 13Hospital, Kissy, Freetown, Sierra Leone 143Partners in Health, Port Loko, Sierra Leone 154World Health Organization, Geneva, Switzerland 165Public Health England Port Loko Laboratory, Port Loko, Sierra Leone 176World Health Organization, Sierra Leone Ebola Response Team, Freetown, Sierra Leone 187Infectious Diseases Institute, Mulago Hospital Complex, Kampala, Uganda 198Department of Medicine, University of Toronto, Toronto, Canada 209GOAL Global, Port Loko, Sierra Leone 2110Liverpool School of Tropical Medicine, Liverpool, UK 2211International Medical Corps, Port Loko, Sierra Leone 2312Department of Critical Care Medicine, Sunnybrook Health Sciences Centre, Toronto, Canada 2413Department of Ophthalmology, Emory University School of Medicine, Atlanta, USA 2514Dignitas International, Toronto, Canada 26 27

Corresponding author: Sharmistha Mishra, MD, PhD. Room 363, 30 Bond Street, Toronto, 28Ontario, M5B 1W8. Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Division of 29Infectious Disease, Department of Medicine, University of Toronto. 30sharmistha.mishra@utoronto.ca 31 32

2

SUMMARY 33

Background Limited data are available on the prevalence and predictors of clinical sequelae in 34

survivors of Ebola virus disease (EVD). The EVD Survivor Clinic in Port Loko, Sierra Leone, 35

has provided clinical care for 603 of 661 survivors living in the district. We did a cross-sectional 36

study to describe the prevalence, nature, and predictors of three key EVD sequelae (ocular, 37

auditory, and articular) in this cohort of EVD survivors. 38

Methods We reviewed available clinical and laboratory records of consecutive patients assessed 39

in the clinic between March 7 and April 24, 2015. We used univariate and multiple logistic 40

regression to examine clinical and laboratory features of acute EVD with the following outcomes 41

in convalescence: new ocular symptoms, uveitis, auditory symptoms and arthralgias. 42

Findings Among 277 survivors (59% female), median age was 29 years (IQR 20–36) and 43

median time from discharge from an EVD treatment facility to first survivor clinic visit was 121 44

days (82–151). Clinical sequelae were common, including arthralgias (n=210, 76%), new ocular 45

symptoms (n=167, 60%), uveitis (n=50, 18%), and auditory symptoms (n=67, 24%). Higher 46

Ebola viral load at acute EVD presentation (as shown by lower cycle thresholds on real-time RT-47

PCR testing) was independently associated with uveitis (adjusted odds ratio [aOR] 3·33, 95% CI 48

1·87–5·91, for every five-point decrease in cycle threshold) and with new ocular symptoms or 49

ocular diagnoses (aOR 3·04, 95% CI 1·87–4·94). 50

Interpretation Clinical sequelae during early EVD convalescence are common and sometimes 51

sight threatening. These findings underscore the need for early clinical follow-up of EVD 52

survivors and urgent provision of ocular care as part of health systems strengthening in EVD-53

affected West African countries. 54

Funding Canadian Institutes of Health Research. 55

3

Copyright © 2015. World Health Organization. Published by Elsevier Ltd/Inc/BV. All rights 56

reserved. 57

58

Lancet Infect Dis 2015 Published Online December 22, 2015: http://dx.doi.org/10.1016/S1473-59

3099(15)00489-2 60

See Online/Comment: http://dx.doi.org/10.1016/S1473-3099(15)00546-0 61

4

INTRODUCTION 62

The Ebola virus disease (EVD) outbreak in west Africa is the largest in history. As of November, 63

2015, over 28 500 EVD cases have been reported with an estimated 15 000 survivors.1 64

Community-led survivor networks have alerted health-care providers to a variety of convalescent 65

symptoms, including vision and hearing loss and arthralgia.2–4 66

67

Understanding of the nature, timing, and prevalence of EVD sequelae remains limited.5–13 68

Disabling sequelae, including ocular, auditory, and arthritic symptoms, have been described in 69

small studies from previous outbreaks.7,9–11,13 In the current outbreak, one qualitative study of 70

100 survivors in Sierra Leone reported blurred or partial loss of vision in convalescence but did 71

not quantify these sequelae.2 Surveys of 105 survivors in Guinea14 and of 81 survivors in Sierra 72

Leone15 noted frequent musculoskeletal pain14,15 and visual problems;15 neither study included a 73

clinical examination of survey participants. No studies from the west African outbreak have 74

examined possible relations between features of acute EVD and the frequency or severity of 75

clinical sequelae. Therefore, we did a cross-sectional study to describe the prevalence, nature, 76

and predictors of three key EVD sequelae (ocular, auditory, and articular) in a large cohort of 77

survivors of EVD in Port Loko district, Sierra Leone. 78

79

METHODS 80

Study Setting 81

By Nov 12, 2015, 1485 EVD cases were reported from the rural district of Port Loko (population 82

572 369), with 661 survivors according to the Sierra Leone Association for Ebola Survivors 83

registry.16 Before Nov 30, 2014, some patients were referred for care in Ebola treatment units 84

5

(ETUs) outside the district, since ETU scale-up was still underway in Port Loko. After Nov 29, 85

2014, 90% of patients with EVD in Port Loko received care at one of three ETUs (Maforki, 86

Mathaska, and Lunsar) in the district. The Port Loko EVD Survivor Care Clinic was established 87

on March 7, 2015, at the Baptist Eye Hospital Lunsar as a clinical partnership between Partners 88

in Health (PIH), the PIH-supported EVD Survivor Network,GOAL Global, International 89

Medical Corps, and Christian Blind Mission under the oversight of the Sierra Leone 90

Ministry of Health and Sanitation (MoHS) District Health Management Team, with technical 91

support from WHO. 92

93

Patient population 94

Survivors of EVD were identified via the MoHS-WHO registry of patients residing in Port Loko 95

District, regardless of where originally treated. All were laboratory confirmed to have EVD 96

through real-time RT-PCR testing on serum and discharged from ETUs after clinical 97

improvement and a negative convalescent real-time RT-PCR. The registry was cross checked 98

against the Sierra Leone Association for Ebola Survivors registry to generate a complete list 99

(appendix). The EVD Survivor Network led community sensitisation regarding the establishment 100

of the clinic. Survivors were systematically contacted by mobile phone by the clinic coordinator 101

according to village of residence. Bus pick-ups were scheduled for each village and patients were 102

assessed in the clinic irrespective of symptoms. We started with villages with resident survivors 103

discharged in the remote past, until all villages were covered. As of Nov 12, 2015, the clinic had 104

assessed 603 survivors of EVD residing in Port Loko at least once. 105

106

Data Collection 107

6

We extracted demographic and clinical data from patient charts (appendix) on the first 277 108

consecutive survivors of EVD assessed in the survivor clinic between March 7, 2015, and April 109

24, 2015. Eight patients were self-referred with ocular symptoms, but all would nevertheless 110

have been identified through the village selections over the study period. Each patient received a 111

clinical assessment and an eye examination, including visual acuity and slit-lamp examination. 112

Patients with ocular symptoms, decreased visual acuity, or abnormalities on slit-lamp 113

examination also received dilated fundoscopic assessment. Clinical data were entered into an 114

electronic database and linked to two other datasets using the EVD laboratory number, and 115

cross-checked with a unique patient identifier, sex, date of acute EVD testing, and patient 116

residence or age. The additional datasets included: EVD surveillance data, which provided 117

symptoms on presentation; and real-time RT-PCR and cycle threshold (an inversely correlated 118

marker of viral load) results on the subset of patients who were diagnosed or cared for in a Port 119

Loko ETU. The real-time RT-PCR assay used for testing changed after Feb 1, 2015 (appendix). 120

After linkage, anonymised data were used for analyses. The study was approved by the Sierra 121

Leone MoHS and Ethics and Scientific Review Committee. 122

123

Data Analysis 124

We used descriptive statistics to report features of acute EVD and clinical symptoms at the first 125

convalescent visit. We used χ² or Fisher’s exact test for categorical data and the t test or 126

Wilcoxon rank-sum test for continuous data to assess the relation between demographics (age, 127

sex), cycle threshold at EVD diagnosis, duration of acute illness (days from symptom onset to 128

the first negative real-time RT-PCR during acute EVD), self-reported symptoms of acute 129

7

EVD, and the absence or presence of each of four key sequelae present at the first convalescent 130

visit: uveitis diagnosed on slit-lamp and dilated fundoscopic examination; new ocular symptom 131

or diagnosis (including uveitis or conjunctivitis); new auditory symptoms (tinnitus, subjective 132

hearing loss, aural fullness); and new arthralgias or diagnosis of arthritis. Symptoms were 133

considered new if onset was either while in the ETU or after discharge. We explored all self-134

reported symptoms of acute EVD in the univariate analyses, excluding individual responses 135

listed as missing or could not recall. 136

137

We used multiple logistic regression to delineate predictors of EVD sequelae. To preserve model 138

parsimony, independent variables were chosen a priori to show severity of acute EVD (duration 139

of acute illness and cycle threshold value at diagnosis17) and acute clinical features based on 140

previously established links to post-infectious immunological sequelae (presence of red eyes 141

[for ocular outcomes only], presence of diarrhoea18). The model was adjusted for age and sex. 142

143

We did a sensitivity analysis to determine if time period of cycle threshold value measurement 144

(before or after real-time RT-PCR assay changed on Feb 1, 2015) affected the final models as an 145

interaction effect. Analyses were done using SAS software, version 9.3 (SAS Institute). 146

147

Role of Funding Source 148

The research was funded in part by the Canadian Institutes of Health Research. The funder had 149

no role in study design, data collection, data analysis, data interpretation, or writing of the report. 150

The study leads (JGM, MJV) and corresponding author (SM) had full access to all the data in the 151

study and had final responsibility for the decision to submit for publication. 152

8

RESULTS 153

Demographic and clinical features during acute EVD in the 277 survivors included in the study 154

are presented in table 1. Just over half (n=163, 59%) were female, and the median age was 29 155

years (IQR 20–36). Median time from ETU discharge to first clinic visit was 121 days (IQR 82–156

151). 157

158

Symptoms and diagnoses at first convalescent clinic visit are presented in table 2. 210 (76%) 159

patients reported arthralgias (joint pain or ache without swelling or evidence of an effusion on 160

examination), predominantly of an oligoarthralgia pattern (one to four joints), with most (n=180, 161

86%) reporting bilateral joint involvement. One patient also had tenosynovitis. 162

163

Patients reported tinnitus (n=56, 20%), aural fullness (n=23, 8%), and subjective hearing loss 164

(n=17, 6%), and 167 patients (60%) reported at least one ocular symptom. The self-reported 165

median time from ETU discharge to onset of symptoms varied according to symptom: articular 166

(1 week [IQR 0–4·3]; n=156), auditory (2 weeks [0–8·6]; n=51), and ocular (2 weeks [0–8·6]; 167

n=119) symptoms. Symptoms began while in the ETU in 52 (25%) of 210 patients with auditory 168

symptoms, 18 (27%) of 67 patients with articular symptoms, and 33 (20%) of 167 patients with 169

ocular symptoms. 170

171

Based on slit-lamp and dilated fundoscopic examination, 50 (18%) patients were diagnosed with 172

uveitis involving 68 eyes (46% anterior, 26% posterior, 3% intermediate uveitis, and 25% 173

panuveitis). Uveitis was predominantly unilateral (64% of cases). Five (10%) patients with 174

uveitis also had early cataracts, and their median age was 29 years (IQR 18–40) years compared 175

9

with 45 years (38–52·5) in the 16 patients with cataracts but without uveitis, although we did not 176

have statistical power to assess for differences. Patients with uveitis developed ocular symptoms 177

a median of 3 weeks (IQR 0·4–8·6; range 0–17·2) after ETU discharge. Each ocular symptom, 178

compared with its absence, was associated with uveitis (appendix). The presence of blurry 179

vision, light sensitivity, or itchy eye was 88·0% sensitive and 50·7% specific for underlying 180

uveitis. Conjunctivitis was present in 31 (11%) of 277 patients. 181

182

Table 3 shows the univariate analyses of differences in demographic, clinical, and laboratory 183

features during acute EVD and sequelae reported at the first convalescent visit. Other than 184

arthralgia, which was more common in older persons, sequelae occurred with similar frequencies 185

in both sexes and across age groups. On additional exploratory analysis, ocular symptoms during 186

acute EVD were associated with subsequent ocular symptoms, but not specifically with uveitis. 187

Acute symptoms such as gingival bleeding, red eyes, and blurry vision were more commonly 188

reported among those with subsequent auditory or articular sequelae. A history of fever during 189

acute EVD was associated with uveitis and with ocular symptoms in convalescence. Lower cycle 190

threshold values at acute EVD presentation were significantly associated with uveitis in 191

convalescence (median 22·4 [IQR 19·5–26·1] for those with uveitis versus 26·8 192

[23·5–29·0] for those without, p<0·0001) but not with arthralgias or auditory symptoms. 193

194

Data on cycle threshold value and duration of acute EVD were available for a subset of 190 195

patients who were either diagnosed or treated in ETUs in Port Loko. The demographic and 196

clinical features of patients with and without laboratory data are shown in the appendix. 197

Patients with missing laboratory data were not significantly different in age, sex, or symptoms 198

10

during acute EVD. 199

200

Table 4 shows the independent predictors of clinical sequelae on multivariable analyses. After 201

adjusting for age, sex, and duration of acute illness, a lower cycle threshold value at acute 202

presentation was independently associated with uveitis (adjusted odds ratio [aOR] 3·33, 203

95% CI 1·87–5·91, for every five-point decrease in cycle threshold) and with new onset ocular 204

symptoms or diagnoses overall (aOR 3·04, 95% CI 1·87–4·94). Neither diarrhoea nor red eyes 205

during acute EVD were associated with uveitis. Nor was diarrhoea during acute EVD associated 206

with arthralgias or auditory sequelae. The association between cycle threshold value and EVD 207

sequelae was not significantly affected by time period (before or after Feb 1, 2015). 208

209

DISCUSSION 210

We found that 50 (18%) patients who survived acute EVD developed uveitis, with ocular 211

symptoms developing as early as during the ETU stay or as late as 17 weeks after discharge. 212

Ebola viral load at the time of EVD diagnosis, which has also been associated with increased 213

mortality in past studies, was the key independent predictor of ocular symptoms and specifically 214

uveitis.17,19,20 215

216

Our findings generate important hypotheses regarding the pathogenesis of EVD sequelae. Ebola 217

virus is rapidly cleared from most body fluids after resolution of acute disease,21–23 but might 218

persist in immunologically privileged sanctuary sites.10,21 For example, viable Ebola virus was 219

isolated from the aqueous humour at high concentrations (cycle threshold of 18·7) in a patient 220

who developed severe sight-threatening uveitis 9 weeks after surviving acute EVD with high 221

11

viraemia.24 The findings from this case and our findings from Sierra Leone support the 222

hypothesis that virus persistence and replication in ocular chambers might play a part in the 223

pathogenesis of uveitis in survivors of EVD.24 Furthermore, the fact that the patient was critically 224

ill during his acute infection is consistent with the hypothesis that severe disease, which 225

generally correlates with level of viraemia,17 might lead to virus persistence and long-term 226

complications. The same survivor had prolonged persistence of virus in the semen (Crozier I, 227

unpublished) and shedding of virus in semen has been reported up to 9 months after acute illness 228

in survivors in Sierra Leone.25 Determination of the early predictors of persistence in semen and 229

its relation to ocular complications would be of interest as a potentially less 230

invasive marker of intraocular persistence. 231

232

Persistent immune activation, rather than direct viral cytopathic effect, has been postulated to 233

lead to some post-EVD sequelae, although the processes are not necessarily mutually 234

exclusive;26 immune activation might be driven by persistent virus replication or delayed antigen 235

clearance in immune-privileged tissues. In our study, there was no association between the level 236

of viraemia at acute EVD presentation and convalescent arthralgias or auditory symptoms, a 237

finding which suggests a potentially different mechanism of non-ocular disease post-EVD. 238

However, we cannot exclude the possibility that the lack of apparent association relates to how 239

outcomes were measured (eg, in this study lack of audiometry) or defined (any joint involvement 240

rather than specific distribution of joints), or to potential differences in levels of viraemia 241

required to penetrate into various sites. IgG antibody titres were significantly higher in 29 242

survivors of EVD with arthralgias than in those without after the 1995 outbreak in Kikwit, 243

12

Democratic Republic of Congo, a finding consistent with persistent immune activation as the 244

pathogenic mechanism.11 Unfortunately, we were not able to do antibody testing on our patients 245

to assess relations between antibody titre and sequelae. 246

247

The prevalence of arthralgias in our series is similar to that noted in a survey of survivors of 248

EVD in Guinea (87%) a median of 103 days into convalescence,14 but was higher than that 249

reported in survivors of EVD at 21 months after the aforementioned Kikwit outbreak (48%)11 250

and up to 29 months after the 2007 Bundibugyo outbreak (25%), perhaps suggesting attenuation 251

over time.13 252

253

Systematic clinical care for survivors of EVD was absent during the early part of the current 254

outbreak due to the overwhelming need to care for those with acute disease. Furthermore, in 255

Sierra Leone, there are just two ophthalmologists and eight mid-level ophthalmological 256

care-providers in the National Eye Health Program. 257

258

Systematic and universal access to ocular care for survivors of EVD is further restricted by lack 259

of equipment (especially slit-lamps) and ocular medications, and by mobility of clinicians and 260

patients. Addressing these barriers holds the potential to provide early diagnoses and treatment to 261

survivors of EVD while strengthening eye care service delivery during health system 262

reconstruction in west Africa. Our findings also signal the need for operational research into the 263

feasibility and safety of syndromic treatment, including the use of topical steroids and 264

cycloplegic agents, for survivors of EVD with clinical presentations consistent with uveitis, 265

especially when slit-lamp examination is not possible. 266

13

267

A recent survey of 105 survivors of EVD in Guinea reported no ocular symptoms or hearing 268

loss.14 The discrepancy with our results might be explained by the fact that the Guinea study only 269

reported vision loss (noted in seven [3%] of our study participants) and that patients 270

were not asked about tinnitus or aural fullness.14 In a survey of 81 survivors of EVD in Kenema, 271

Sierra Leone, up to at least 4 months after resolution of acute disease, 42% reported visual 272

problems.15 The variability in prevalence of sequelae reinforces the need for systematic and 273

harmonised clinical assessments (including physical and ocular examination) and data collection 274

and analyses across west Africa. 275

276

Although various case reports and surveys have been published,14,15,24,27,28 the data reported here 277

comprise the largest systematic study of survivors of EVD from the 2014–15 outbreak. 278

Furthermore, unlike studies from previous clinical cohorts, the clinical data are drawn from a 279

more representative sample of survivors, with all but eight (3%) of the participants (who self-280

presented with ocular complaints) drawn from systematic inclusion of villages and assessed 281

irrespective of symptoms. Furthermore, protocol-defined clinical charting and documentation 282

provided structured data. Because all patients underwent a slit-lamp examination and, when 283

indicated, a dilated fundoscopic examination, the prevalence of uveitis noted is a reliable 284

estimate, with minimal selection and measurement bias. However, since our data are derived 285

from baseline clinic visits done at a range of periods after the onset of disease, the frequencies of 286

the symptoms recorded might not reflect those that would be noted if clinic visits were scheduled 287

at uniform times after recovery from acute EVD. 288

289

14

There are important limitations to this study. First, its cross-sectional nature prohibits 290

measurement of the true incidence of EVD sequelae. Second, we did not have a comparison 291

group of patients not infected with Ebola virus, and thus cannot infer the relative risk of the 292

sequelae against background rates from other aetiologies of arthralgias, auditory, or ocular 293

symptoms endemic to west Africa. Third, the clinic was not equipped for audiometry to 294

objectively assess hearing. Fourth, symptoms during acute EVD were drawn from surveillance 295

data (collected at the time of acute infection) and via symptom recall during convalescence, 296

which might have resulted in recall bias. The proportion of patients with fever and diarrhoea 297

during acute EVD noted in our study is similar to the proportion reporting fever (84–89%) and 298

diarrhoea (51–62%) in other smaller clinical cohorts of acute EVD.29–31 However, red eyes and 299

severe symptoms such as bleeding (including gingival bleeding) were more common compared 300

with data collected at the time of acute EVD diagnosis in other studies.29–31 The higher frequency 301

of certain symptoms might also reflect new symptoms developing during acute illness but that 302

were not present at the time of diagnosis. Fifth, the denominator of EVD survivors in the district 303

might vary with inter-district migration, or willingness of survivors to disclose their current 304

district of residence to the survivor network. Finally, laboratory data were missing for a subset of 305

patients, although the primary outcomes (specifically uveitis) and other patient characteristics 306

were similar to those for whom laboratory data were available. 307

308

Further research is needed to understand the pathogenesis of various EVD sequelae and to 309

optimise treatment. The largest EVD outbreak in history has logically led to the largest number 310

of survivors of EVD in history. These findings emphasise the importance of ongoing clinical 311

follow-up and care of all patients, starting at discharge from an ETU, and underscore the urgent 312

15

need for the greater provision of ocular care as part of the strengthening of health systems in 313

west Africa. 314

315

RESEARCH IN CONTEXT 316

Evidence before this study 317

A MEDLINE search on the prevalence of post-Ebola virus disease (EVD) clinical sequelae using 318

search terms “Ebola” and “survivor OR sequelae OR convalescen*” yielded 227 unique citations 319

published by Nov 12, 2015. There were no language restrictions. Excluding case reports, 320

commentaries, and expert reviews, nine studies (case series, cohorts, and cross-sectional surveys) 321

provided clinical information post-Ebola sequelae, including three from the current west Africa 322

outbreak. Only six studies quantify the prevalence of clinical sequelae, of which four explored 323

early clinical sequelae in 240 survivors of Ebola within 3–4 months of convalescence. To date, 324

measurement of clinical sequelae was based on a clinical examination in only 57 patients (from 325

two studies of previous outbreaks), and only four patients in these studies received a complete 326

ophthalmological examination (visual acuity, slit-lamp, dilated fundoscopy). The largest (n=105) 327

published study of early EVD clinical sequelae was based on self-reported symptoms without a 328

clinical examination, with restricted questions on ocular and auditory symptoms that were self-329

reported in none of the 105 survivors of EVD. None of the studies examined the predictors of 330

developing early or late EVD clinical sequelae. 331

332

Added value of this study 333

This study’s systematic clinical examination of EVD sequelae includes the largest representative 334

sample of west African survivors of EVD from the 2014–15 outbreak. Unlike studies from 335

previous clinical cohorts, the clinical and laboratory data were drawn from a more representative 336

16

sample, with detailed and protocol-defined clinical charting and clinical examination, and 337

provide information on the timing of sequelae. All patients underwent a slit-lamp examination, 338

and as indicated, a dilated fundoscopic examination, such that the prevalence of uveitis is a 339

reliable estimate with minimal selection and measurement bias. This study is also the first to 340

examine the clinical and laboratory predictors of EVD clinical sequelae in convalescence. 341

342

Implications of the available evidence 343

These findings signal an immediate need to systematically provide early clinical follow-up for all 344

survivors of EVD with particular attention paid to the potential for ocular complications. Further 345

research is needed to understand the pathologies underlying the various EVD sequelae. 346

347

CONTRIBUTORS 348

JGM, MJV, SMi, JCC, LM, SMa, and AR conceived of and designed the study. AKC, JGM, 349

DGB, and SMi did the literature search. JGM, JCC, SMi, LM, SMa, DEP, KD, SC, JS, SY, 350

RAF, AKC, and MJV developed the data collection tools. JGM, YM, APK, JCC, DEP, KD, SC, 351

APS, SMi, LM, TB, AR, and VP collected the data. RP, SMi, AKC, and RAF did the data 352

analysis. All authors contributed equally to the data interpretation. JGM, SMi, and AKC wrote 353

the first version of the manuscript. All authors contributed equally to the critical review and 354

editing of the manuscript. 355

356

DECLARATION OF INTERESTS 357

TO’D, AKC, and SMi report personal fees as clinical consultants from WHO during the conduct 358

of the study. VP, AR, and VW report personal fees from International Medical Corps, financed 359

through donor funding from USAID/OFDA and Children’s Investment Fund Foundation, during 360

17

the conduct of the study. JCC, DEP, KD, SC, CK, and LM are employed by Partners in Health, 361

which received funding from USAID and the Paul Allen Foundation to support the clinical care 362

and services provided in the Ebola survivor clinic in Port Loko, Sierra Leone. All other 363

authors declare no competing interests. 364

365

ACKNOWLEDGEMENTS: 366

We thank the patients in the Port Loko Ebola Survivor Clinic. We acknowledge the coordination 367

and leadership of the PIH-supported Port Loko EVD Survivor Network in the outreach and clinic 368

activities (Jamil Bangura, Francis Bangura, Sherry Bangura, Michael Drasher, Ibrahim Kargbo, 369

Suwad Kamara, Bai S. Koroma, Alieu Kargbo, Ibrahim Sesay, Marie M. Sesay, Humu Turay, 370

Gabriel W. Schlough). PIH provided the clinical leadership and contributed financially to clinical 371

care and associated services at the clinic. The clinic was collectively conceived of, designed, and 372

initiated by John Welch (Partners in Health [PIH]), Margarite D’Amelio (GOAL Global), 373

Audrey Rangel (International Medical Corps), John Mattia (Baptist Eye Hospital, Christian 374

Blind Mission, Sierra Leone National Eye Health Program), Matthew Vandy (Sierra Leone 375

National Eye Health Program) in partnership with the Port Loko Ministry of Health and 376

Sanitation District Health Management Team, the Port Loko PIH-supported EVD Survivor 377

Network, and with technical assistance from the WHO Ebola Survivor Support Network (WHO-378

ESSN). We thank the staff at the Baptist Eye Hospital in Lunsar (Ibiana D. Williams, John S. 379

Kabba, Saffiatu A. Koroma, John Deen Sankoh, Amadu P. Kamara, Paul L. Kamara, Kolleh H. 380

Kanu, Francis Ndanema, Haja S. Mansaray, Abu Koroma, Agnes I. Kamara, Esther Turay, 381

Zainab Bangura, Abu Turay, Momoh Kargbo, Osman Nelson, Idrissa Tarawalie, Sheka 382

Tarawalie, Apha Jalloh, Isatu O. Kamara, Kadaitu Gbla, Amidu Kahina). We thank the clinical 383

18

and monitoring and evaluation staff from PIH: Edward Vandi, Andrew Sesay, Joanna Balza, 384

Nick Sarchet, Gayle Rebovich, Musa Sillah, Jacob Bumstead, Janaki Kuruppu, Lisa Benaron, 385

Alimamy Kamara, Abdul A. Kamara. We thank all health-care providers from GOAL Global 386

(Jennifer Keys, Carlos Mutingamo Junior Dibungu, Ahmed Aldikhari, Jonjoe Sampson, Yassin 387

Jalloh, Alex Kamara, Mary Sesay, Foday Salam, Desmond Kain, Gibril Turay, Mohamed Potho 388

Kamara, John Kamara, Thomas Joana, Olive Barlay), the Danish Foreign Medical Team 389

(Christian Christoffersen, Andreas Lundh, Louise Lyckhage, Jens Veir, Rikke Thoft Nielsen, 390

Pernille Ravn, Nina Friis-Moeller, Sine Janum, Lothar Weise, Rune Lykke, Trine Petersen, 391

Sarah Marlow, Katrine Lind, Mia Bull, Charlotte Nordsted), and the International Medical Corps 392

(Samuel Kafoe, Brima Sesay, Kashif Islam Siddiqui, Carrie Trabka, Steven Whiteley). 393

394

We acknowledge the leadership and support of the District Medical Officer, Dr. Tom Sesay. We 395

thank Dr. Boris Pavlin, Dr. Frank Grenier, and Dr. Richard Hudspeth (WHO Port Loko) for their 396

support with the clinical charting forms and clinic initiation. We thank Dr. Lowell Gess (Kissy 397

Eye Hospital, Freetown) for ophthalmologic expertise into the early management of ocular 398

complications of EVD, and advocating for urgent ocular care for EVD survivors. We thank Dr. 399

Jan Hajek (University of British Columbia, Canada) and Dr. Neill Adhikari (University of 400

Toronto, Canada) for reading the manuscript and providing critical comments. 401

402

We acknowledge the entire staff of the Public Health England Port Loko laboratory for carrying 403

out the Ebola RT-PCR diagnostic tests and Neill Keppie for his help with Public Health England 404

database queries. We thank the leadership of the Department for International Development for 405

their support of Public Health England laboratory activities. 406

19

407

SMi is supported by a Canadian Institutes of Health Research New Investigator Award. 408

20

Table 1. Demographic and clinical features during acute Ebola virus disease illness among 409277 survivors. 410 Number of survivors (%)

Sex

Female 163 (59%)

Male 114 (41%)

Age group

<5 years 6 (2%)

5-20 years 64 (23%)

21-40 years 158 (57%)

>40 years 49 (18%)

Self-reported acute EVD clinical features1

Fever 255 (92%)

Diarrhoea2 211 (77%)

Eye redness3 207 (75%)

EVD = Ebola virus disease 411

1At any time during acute EVD illness (before and within Ebola facility, including the Ebola 412

treatment unit), self-reported; fever included history of fever within the 2 days before or 413

temperature >38·0°C at the time of EVD case-investigation by the surveillance team or Ebola 414

facility admission. 415

2Missing data in 1 patient (cannot recall and not answered in the surveillance data) 416

3Missing data on 2 patients (cannot recall and not answered from the surveillance data) 417

418

419

420

421

21

Table 2. Ebola virus disease sequelae at first convalescent clinic visit in 277 survivors. 422 Number of

survivors (%)

Arthralgias (symptom onset during ETU or after discharge) 210 (76%)

Any auditory symptoms (tinnitus, aural fullness, hearing loss)

that started during ETU or after discharge

67 (24%)

Ocular symptoms (onset during ETU or after discharge) 167 (60%)

Blurry vision 104 (38%)

Light sensitivity 86 (31%)

Itchy eye 86 (31%)

Tearing 79 (29%)

Pain 72 (26%)

Foreign body sensation 68 (25%)

Floaters 46 (17%)

Redness 46 (17%)

Flashes of light 43 (16%)

Dry eye 39 (14%)

Burning sensation 29 (11%)

Loss of vision 7 (3%)

Uveitis diagnosed on slit-lamp and dilated fundoscopic

examination* (68 eyes involved among 50 patients)

50 (18%)

Anterior uveitis 31 (46%)

Posterior uveitis 18 (26%)

Intermediate uveitis 2 (3%)

Panuveitis 17 (25%)

ETU = Ebola treatment unit. *A patient could have >1 type of uveitis if both eyes were involved. 423

22

Table 3. Features of acute Ebola virus disease illness associated with convalescent symptoms and diagnoses at first survivor clinic visit. Uveitis Ocular symptom or ocular diagnoses1 Auditory symptoms2 Arthralgias2

Yes

(N=50)

No

(N=227)

p Yes

(N=167 )

No

(N=110 )

p Yes

(N=67 )

No

(N=210 )

p Yes

(N=210 )

No

(N=67 )

p

Sex 0⋅08 0⋅08 0⋅08 0⋅90

Female 35(70%) 128 (56%) 105 (63%) 58 (53%) 40 (60%) 123 (59%) 124 (59%) 39 (58%)

Male 15 (30%) 99 (44%) 62 (37%) 52 (47%) 27 (40%) 87 (41%) 86 (41%) 28 (42%)

Age-group (years) 0⋅343 0⋅313 0⋅183 0⋅0043

< 5 0 (0%) 6 (3%) 3 (2%) 3 (3%) 0 (0%) 6 (3%) 3 (1%) 3 (4%)

5-20 10 (20%) 54 (24%) 33 (20%) 31 (28%) 15 (22%) 49 (23%) 40 (19%) 24 (35%)

21-40 34 (68%) 124 (55%) 98 (59%) 60 (55%) 35 (52%) 123 (59%) 124 (59%) 34 (51%)

>40 6 (12%) 43 (19%) 33 (20%) 16 (15%) 17 (25%) 32 (15%) 43 (20%) 6 (9%)

Acute EVD clinical

features4

Fever5 50 (100%) 205 (90%) 0⋅023 161 (96%) 94 (85%) 0⋅001 62 (93%) 193 (92%) 0⋅87 194 (92%) 61 (91%) 0⋅72

Fatigue 49 (98%) 221 (98%) 1⋅003 163 (98%) 107(98%) 1⋅003 64 (96%) 206 (99%) 0⋅163 206 (98%) 64 (97%) 0⋅633

Headache 48 (96%) 206 (91%) 0⋅383 157 (94%) 97 (89%) 0⋅13 60 (90%) 194 (93%) 0⋅39 196 (93%) 58 (88%) 0⋅15

Loss of appetite 47 (94%) 195 (86%) 0⋅13 149 (89%) 93 (85%) 0⋅34 61 (91%) 181 (87%) 0⋅33 183 (87%) 59

(89%)

0⋅62

Nausea or vomiting 40 (80%) 173 (77%) 0⋅60 134 (80%) 79 (72%) 0⋅13 51 (76%) 162 (78%) 0⋅81 163 (78%) 50 (76%) 0⋅75

Abdominal pain 34 (68%) 158 (70%) 0⋅79 114 (68%) 78 (72%) 0⋅56 49 (73%) 143 (68%) 0⋅47 148 (70%) 44 (67%) 0⋅56

Diarrhoea 41 (82%) 170 (75%) 0⋅31 130 (78%) 81 (74%) 0⋅561 50 (75%) 161 (77%) 0⋅69 160 (76%) 51(77%) 0⋅86

Blood in the stool 14 (28%) 61 (27%) 0⋅89 42 (25%) 33 (31%) 0⋅33 19 (29%) 56 (27%) 0⋅75 62 (30%) 13 (20%) 0⋅11

23

Myalgias or arthralgias 47 (94%) 203 (90%) 0⋅443 155 (93%) 95 (87%) 0⋅12 65 (97%) 185 (89%) 0⋅04 193 (92%) 57 (86%) 0⋅18

Sore throat 14 (28%) 92 (41%) 0⋅09 114 (68%) 56 (51%) 0⋅005 14 (21%) 92 (44%) 0⋅0007 75 (36%) 31 (47%) 0⋅10

Hiccups 18 (26%) 86 (36%) 0⋅97 57 (34%) 43 (40%) 0⋅.37 32 (48%) 68 (33%) 0⋅02 78 (37%) 22 (33%) 0⋅57

Red eyes 37 (74%) 170 (76%) 0⋅82 134 (80%) 73 (68%) 0⋅02 57 (85%) 150 (72%) 0⋅03 165 (79%) 42 (65%) 0⋅02

Blurry vision 27 (54%) 136 (60%) 0⋅42 108 (65%) 55 (50%) 0⋅02 47 (70%) 116 (56%) 0⋅03 131 (62%) 32 (49%) 0⋅05

Bleeding gums 11 (22%) 68 (30%) 0⋅25 50 (30%) 29 (27%) 0⋅55 27 (40%) 52 (25%) 0⋅02 67 (32%) 12 (18%) 0⋅03

Miscarriage 2 (4%) 10 (4%) 1⋅003 7 (4%) 5 (5%) 1⋅003 3 (4%) 9 (4%) 1⋅003 8 (4%) 4 (6%) 0⋅493

Cycle threshold values

at EVD diagnosis

(N=190)6,7, Median

(IQR)

22.4

(19.5-26.1)

26⋅8

(23⋅5-29⋅0)

<0⋅0001 25

(22⋅0-28⋅0)

27⋅7

(24⋅5-29⋅5)

<0⋅0001 26⋅5

(23⋅5-28⋅1)

26

(22⋅5-28⋅8)

0⋅96 26⋅5

(23⋅2-28⋅9)

24⋅1

(22⋅1-28⋅6)

0⋅15

Duration of acute EVD

(N=190) 6,7, Median

(IQR)

14 (9-19) 12 (9-17) 0⋅185 13 (9-16) 12 (8-18) 0.54 13 (9-18) 13 (9-17) 0⋅87 12 (9-17) 14 (9-19) 0.15

Data are n (%) or median (IQR) unless otherwise stated. EVD=Ebola virus disease.

1Symptoms began during ETU stay or after discharge; diagnoses includes uveitis or conjunctivitis

2Symptoms began during ETU stay or after discharge

3Fisher’s exact test

4At anytime during acute EVD illness (before and within Ebola facility, including ETU), self-reported. Fever included history of fever within the 2 days prior to Ebola facility

admission or temperature >38.0°C at the time of EVD case-investigation by the surveillance team. Acute EVD symptom data were missing in 32 of 4,155 responses (0.1%).

5All patients with uveitis had a history of fever during acute EVD

6A subset of patients had data on cycle-threshold and duration of acute EVD

7Wilcoxon rank-sum 2-sided test was used because the distribution of the data were skewed

24

Table 4. Features of acute Ebola virus disease associated with sequelae: multivariable analyses Uveitis Any ocular symptom of

diagnoses1

Any auditory symptom2 Arthralgias2

OR (95% CI) p OR (95% CI) p OR (95% CI) p OR (95% CI) p

Female (vs. male) 1⋅73 (0⋅73-4⋅09) 0⋅21 1⋅57 (0⋅81-3⋅07) 0⋅18 1⋅09 (0⋅53-2⋅23) 0⋅82 0⋅94 (0⋅47-1⋅87) 0⋅86

Age in years3 1⋅08 (0⋅92-1⋅26) 0⋅34 1⋅21 (1⋅07-1⋅37) 0⋅002 1⋅06 (0⋅94-1⋅20) 0⋅34 1⋅07 (0⋅95-1⋅21) 0⋅27

Self-reported acute EVD clinical

features4

Diarrhoea 0⋅75 (0⋅28-1⋅99) 0⋅56 0⋅91 (0⋅41-1⋅99) 0⋅81 0⋅90 (0⋅39-2⋅07) 0⋅81 0⋅81 (0⋅35-1⋅88) 0⋅62

Red eyes 0⋅74 (0⋅30-1⋅79) 0⋅50 1⋅52 (0⋅73-3⋅17) 0⋅26 - - - -

Cycle threshold values at EVD

diagnosis (N=190)5,6

3⋅33 (1⋅87-5⋅91) <0⋅0001 3⋅04 (1⋅87-4⋅94) <0⋅0001 1⋅06 (0⋅68-1⋅66) 0⋅80 0⋅84 (0⋅55-1⋅28) 0⋅41

Duration of acute EVD (N=190)5,7 1⋅11 (0⋅84-1⋅46) 0⋅48 1⋅01 (0⋅80-1⋅27) 0⋅93 1⋅00 (0⋅78-1⋅28) 0⋅97 0⋅84 (0⋅67-1⋅05) 0⋅12

EVD (Ebola virus disease). OR (odds ratio). 1Symptoms began during ETU stay or after discharge; diagnoses include uveitis or conjunctivitis. 2Symptoms began during ETU stay or after discharge. 3For every 5-year increase in age 4At anytime during acute EVD illness (before and within Ebola facility, including ETU), self-reported. Fever included history of fever within the 2 days prior to Ebola facility

admission or temperature >38.0°C at the time of EVD case-investigation by the surveillance team. 5Restricted to subset of the population with cycle threshold values and data on duration of acute EVD. 6For every 5-point decrease in the cycle threshold value. 7For every increase in 5 days (from symptom onset to first negative EVD RT-PCR during acute EVD infection).

25

REFERENCES

1.WHO. Ebola Situation Report 11 November 2015. Geneva, Switzerland: World Health Organization, 2015.

http://apps.who.int/ebola/current-situation/ebola-situation-report-11-november-2015 (accessed Nov 12, 2015)

2. Lee-Kwan SH, DeLuca N, Adams M, et al. Support services for survivors of Ebola virus disease - Sierra Leone,

2014. MMWR Morb Mortal Wkly Rep. 2014; 63:1205-6.

3. WHO. Sierra Leone: helping the Ebola survivors turn the page. Freetown, Sierra Leone: World Health

Organization, 2014. http://www.who.int/features/2014/post-ebola-syndrome/en/ (accessed July 7, 2015).

4. WHO. A story of Ebola survival and return. Freetown, Sierra Leone: World Health Organization, 2014.

http://www.who.int/tdr/news/2014/ebola-survival-return/en/ (accessed July 7, 2015)

5. Pattyn SR, ed. Ebola Virus Haemorrhagic Fever. Amsterdam, Netherlands: Elsevier/North-Holland Biomedical

Press, 1977.

6. Ebola haemorrhagic fever in Sudan, 1976. Report of a WHO/International Study Team. Bull World Health Organ

1978; 56:247-70.

7. De Roo A, Ado B, Rose B, Guimard Y, Fonck K, Colebunders R. Survey among survivors of the 1995 Ebola

epidemic in Kikwit, Democratic Republic of Congo: their feelings and experiences. Trop Med Int Health. 1998;

3:883-5.

8. Okware SI, Omaswa FG, Zaramba S, et al. An outbreak of Ebola in Uganda. Trop Med Int Health. 2002; 7:1068-

75.

9. Bwaka MA, Bonnet MJ, Calain P, et al. Ebola hemorrhagic fever in Kikwit, Democratic Republic of the Congo:

clinical observations in 103 patients. J Infect Dis. 1999; 179 Suppl 1:S1-7.

10. Kibadi K, Mupapa K, Kuvula K, et al. Late ophthalmologic manifestations in survivors of the 1995 Ebola virus

epidemic in Kikwit, Democratic Republic of the Congo. J Infect Dis. 1999; 179 Suppl 1:S13-4.

11. Rowe AK, Bertolli J, Khan AS, et al. Clinical, virologic, and immunologic follow-up of convalescent Ebola

hemorrhagic fever patients and their household contacts, Kikwit, Democratic Republic of the Congo. Commission

de Lutte contre les Epidemies a Kikwit. J Infect Dis .1999; 179 Suppl 1:S28-35.

12. Wendo C. Caring for the survivors of Uganda's Ebola epidemic one year on. Lancet. 2001; 358:1350.

13. Clark DV, Kibuuka H, Wakabi S, et al. Long-term sequelae after Ebola virus disease in Bundibugyo, Uganda:

a retrospective cohort study. Lancet Inf Dis. 2015; 15:905-12

26

14. Qureshi AI, Chughtai M, Loua TO, et al. Study of Ebola virus disease survivors in Guinea. Clin Infect Dis

2015; 61:1035-42

15. Nanyonga M, Saidu J, Ramsay A, Shindo N, Bausch DG. Sequelae of Ebola Virus Disease, Kenema District,

Sierra Leone. Clin Inf Dis. 2015; Published online: Sep 8, 2015, http://doi: 10.1093/cid/civ795

16. Ministry of Health and Sanitation, Sierra Leone. Ebola Viral Disease Situation Report 12, Nov 2015. Freetown,

Sierra Leone: Government of Sierra Leone.

17. Ksiazek TG, Rollin PE, Williams AJ, et al. Clinical virology of Ebola hemorrhagic fever (EHF): virus, virus

antigen, and IgG and IgM antibody findings among EHF patients in Kikwit, Democratic Republic of the Congo,

1995. J Infect Dis. 1999; 179 suppl 1:S177-87.

18. Porter CK, Choi D, Riddle MS. Pathogen-specific risk of reactive arthritis from bacterial causes of foodborne

illness. J Rheumatol. 2013; 40:712-4.

19. Fitzgerald G, Vogt F, Moi Gbabai OB et al. The Contribution of Ebola Viral Load at Admission and Other

Patient Characteristics to Mortality in a Médecins Sans Frontières Ebola Case Management Centre, Kailahun, Sierra

Leone, June–October 2014. J Infect Dis. 2015; 212:1752-1758.

20. Towner JS, Rollin PE, Bausch DG. Rapid diagnosis of Ebola hemorrhagic fever by reverse transcription-PCR

in an outbreak setting and assessment of patient viral load as a predictor of outcome. J Virol. 2004; 78:4330-4.

21. Bausch DG, Towner JS, Dowell SF, et al. Assessment of the risk of Ebola virus transmission from bodily fluids

and fomites. J Infect Dis. 2007; 196 Suppl 2:S142-7.

22. Rodriguez LL, De Roo A, Guimard Y, et al. Persistence and genetic stability of Ebola virus during the outbreak

in Kikwit, Democratic Republic of the Congo, 1995. J Infect Dis. 1999; 179 Suppl 1:S170-6.

23. Formenty P, Leroy EM, Epelboin A, et al. Detection of Ebola virus in oral fluid specimens during outbreaks of

Ebola virus hemorrhagic fever in the Republic of Congo. Clin Infect Dis. 2006; 42:1521-6.

24. Varkey JB, Shantha JG, Crozier I et al. Persistence of Ebola Virus in Ocular Fluid during Convalescence. N

Engl J Med. 2015; 372:2423-7.

25. Deen GF, Knust B, Broutet N, et al. Ebola RNA persistence in semen of Ebola virus disease survivors –

Preliminary report. N Engl J Med. 2015; Published online: Oct 14, 2015, http://doi: 10.1056/NEJMoa1511410

26. McElroy AK, Akondy RS, Davis CW, et al. Human Ebola virus infection results in substantial immune

activation. Proc Natl Acad Sci USA. 2015; 112:4719-24.

27

27. Mora-Rillo M, Arsuaga M, Ramírez-Olivencia G. Acute respiratory distress syndrome after convalescent

plasma use: treatment of a patient with Ebola virus disease contracted in Madrid, Spain. Lancet Resp Med. 2015;

3:554-62.

28. Mohammed A, Sheikh TL, Gidado S, et al. An evaluation of psychological distress and social support of

survivors and contacts of Ebola virus disease infection and their relatives in Lagos, Nigeria: a cross sectional study--

2014. BMC Public Health. 2015; 15:824

29. Bah EI, Lamah MC, Fletcher T, et al. Clinical presentation of patients with Ebola virus disease in Conakry,

Guinea. N Engl J Med. 2015; 372:40-7

30. Schieffelin JS, Shaffer JG, Goba A, et al. Clinical illness and outcomes in patients with Ebola in Sierra Leone. N

Engl J Med. 2014; 371:2092-100.

31. Qin E, Jingfeng B, Zhao M, et al. Clinical features of patients with Ebola virus disease in Sierra Leone. Clin

Infect Dis. 2015; 61:491-5