Sensitive poliovirus detection using nested PCR and nanopore sequencing: a prospective validation study

Timely detection of outbreaks is needed for poliovirus eradication, but gold standard detection in the Democratic Republic of the Congo takes 30 days (median). Direct molecular detection and

nanopore sequencing (DDNS) of poliovirus in stool samples is a promising fast method. Here we report prospective testing of stool samples from suspected polio cases, and their contacts, in

the Democratic Republic of the Congo between 10 August 2021 and 4 February 2022. DDNS detected polioviruses in 62/2,339 (2.7%) of samples, while gold standard combination of cell culture,

quantitative PCR and Sanger sequencing detected polioviruses in 51/2,339 (2.2%) of the same samples. DDNS provided case confirmation in 7 days (median) in routine surveillance conditions.

DDNS enabled confirmation of three serotype 2 circulating vaccine-derived poliovirus outbreaks 23 days (mean) earlier (range 6–30 days) than the gold standard method. The mean sequence

similarity between sequences obtained by the two methods was 99.98%. Our data confirm the feasibility of implementing DDNS in a national poliovirus laboratory.

Despite substantial progress made by the Global Polio Eradication Initiative (GPEI), since inception in 1988, poliomyelitis remains a major public health problem in countries with low

vaccination coverage. Mass vaccination campaigns with oral poliovirus vaccine (OPV) are used during poliovirus outbreaks to stop transmission. However, a combination of slow shipping of

stool samples, time-consuming virus isolation using cell culture and insufficient sequencing capacity delay outbreak responses and reduce the impact of mass vaccination campaigns1,2,3.

In August 2020, the African Region was declared to have interrupted the transmission of wild poliovirus (WPV)4. Vaccination with OPV has resulted in outbreaks of circulating vaccine-derived

poliovirus (cVDPV), which occurs by reversion of attenuating mutations in the live-vaccine strain. Live-vaccine strains are shed in faeces following vaccination and can spread in

under-immunized populations, with the attenuating mutations being lost over time5,6. Serotype 2 cVDPV (cVDPV2) epidemics in young children plague Africa, and Western Asia, in this post-WPV

era. In 2020, 959 cases of paralysis caused by cVDPV2 were reported in 27 countries, including 21 countries in Africa7; in 2021, 692 cases caused by cVDPV2 and 20 cases by serotype 1 cVDPV

were reported globally, mainly in Africa including Nigeria and the Democratic Republic of the Congo (DRC)8. In 2022 at least 843 cases of VDPV were reported, 502 of which were in the DRC8.

In DRC, 10 years after the last case of WPV, there has been almost continual circulation of cVDPV2 as a result of emergence of cVDPV2 following the use of serotype 2 OPV in response to

existing outbreaks. Responses are hampered by inadequate surveillance and lengthy times before outbreaks are confirmed. Poliovirus surveillance is based on the collection of stool samples

from children with acute flaccid paralysis (AFP) and their contacts, and on environmental (sewage) sampling. Effective poliovirus surveillance relies on high-quality sample collection and

laboratory testing. Stool collected from an AFP case is considered adequate for testing if two stools are collected 48 h apart, within 2 weeks of onset of paralysis, and arrive by cold chain

with proper documentation. In DRC, the proportion of AFP cases with inadequate stool sample collection was 23% in 2018 (ref. 8). Additionally, logistical challenges in sample shipment to

the laboratory, in laboratory testing of the samples and in international shipment (to South Africa) for sequencing cause delayed detection of poliovirus outbreaks. Case numbers from an

outbreak have been estimated to increase by approximately 12% (95% credible interval 5–21%) per additional week9 (average of data for the African Region) owing to these logistical problems.

The World Health Organization (WHO) identified delays in detection as one of the major challenges facing the polio eradication programme10.

The DRC comprises 2,345,000 km² but has just one WHO-accredited laboratory, at the Institut National de Recherche Biomédicale (INRB) in Kinshasa, that is responsible for country-wide

biological diagnosis of poliovirus. The INRB uses a sensitive and standardized WHO detection protocol that combines cell culture with intratypic differentiation (ITD) quantitative PCR

(qPCR). Sequencing of the poliovirus VP1 capsid region is carried out at a separate laboratory in the Republic of South Africa, with a VP1 sequence required to both confirm poliovirus

detection, or cases and to distinguish cVDPV from vaccine strains.

The GPEI is currently considering which approaches to use to achieve polio eradication in the last two WPV-endemic countries, Afghanistan and Pakistan, and to combat outbreaks of cVDPV in

four of WHO’s six geographical regions3. The WHO Polio Eradication Strategy 2022–2026 (ref. 3) committed to improvements in detection and response. This includes direct detection of

poliovirus in stool samples, thereby removing the need for the cell-culture-based detection algorithm according to the worldwide poliovirus containment aims11, and shifting of poliovirus

testing and sequence analysis to country level.

These improvements in detection and response could be achieved by implementation of a direct molecular detection and nanopore sequencing (DDNS) method12. DDNS combines fast, direct detection

from stool samples with on-site sequencing, avoiding international transport of samples and enabling quick response to outbreaks9. It could be implemented in any laboratory already using

PCR, including INRB, in which Illumina and nanopore sequencing have been used for Ebolavirus, measles, monkeypox and severe acute respiratory syndrome coronavirus 2 (refs. 13,14).

In this Article, to validate DDNS implementation so that it can be considered as a recommended method by the WHO Global Polio Laboratory Network (GPLN)15, we undertook a prospective study in

the DRC to evaluate application of the DDNS protocol and compared it with gold standard cell-culture methods for poliovirus surveillance. Here we report the sensitivity and specificity of

DDNS compared with cell culture, sequencing accuracy, time taken in the lab and associated cost data.

Stool samples were tested in parallel using both the DDNS and the gold standard assay. A total of 2,339 prospective stool samples, from 1,159 AFP cases (each yielding 1 or 2 samples) and 62

case contacts or community samples (each 1 sample), were processed using 26 nanopore sequencing runs in a 141 day period, averaging one sequencing run every 5.4 days. DDNS identified 62

samples (2.7% of total samples) as positive for poliovirus, with 36 cVDPV2 (1.58%), 5 Sabin serotype 1 (0.30%), 19 Sabin serotype 3 (0.90%) and 2 that contained serotypes 1 and 3 Sabin

poliovirus (0.09 %) (Table 1). The gold standard assay identified polioviruses in 51 samples, of which 31 samples were serotype 2 VDPV (1.33%), 4 were Sabin serotype 1 (0.17%) and 16 were

Sabin serotype 3 (0.68%).

The sensitivity and specificity of detection for each poliovirus type for DDNS or the current gold standard assay is presented in Table 2. cVDPV2 detected by either method were not

contamination because sequences differed from those of other samples (as shown in Supplementary Fig. 1). The sensitivity and specificity of the two methods did not differ significantly

(Fisher’s exact test).

Two stool samples were available for 1,118 AFP cases, with 37 cases positive for poliovirus by either method. Eighteen cases had full concordance between both methods with both samples

testing positive (Supplementary Table 1). There were no cases where both samples tested positive by the gold standard assay and yielded no positive DDNS result, whereas in nine cases with

positive DDNS results no poliovirus was detected by the gold standard assay.

A single sample or pair of samples were available for 1,159 AFP cases. The sensitivity and specificity of detection were calculated for each AFP case (Table 3), and for only AFP cases where

two stools were available (n = 1,118 cases; Supplementary Table 2).

During this study period, 27 samples containing VDPV2 had the VP1 region sequenced using both diagnostic methods. Only samples of programmatic importance (where vaccination response may be

required; all serotype 2 viruses and any suspected vaccine-derived and wild-type polioviruses) are sequenced following cell culture whereas DDNS produces a sequence for positive samples

without requiring additional sequencing elsewhere. For these 27 samples a median of 6 days was required between case onset and sample collection (range 2–21 days) and a further median 6 days

was required between sample collection and arrival of samples at the sequencing laboratory (range 2–27 days). The time from receipt in the sequencing laboratory to a VP1 sequence took a

median of 30 days (range 21–41 days) via the standard algorithm, including a median of 8 days (range 4–22 days) required for shipment between the virus isolation and sequencing lab, while

DDNS was significantly quicker (P