Identification of Mf positive participants eligible for direct IVM treatment and follow-up
In the Bafia HD (Table 1), a total of 388 participants composed of 311 adults (≥20 years) and 77 children (5-19 years) were recruited for screening by microscopy. Most of the adults are farmers (236 participants). The gender and age group distribution in each community were determined (Table 1, S2 Table).
Of 388 individuals who provided skin biopsies (Table 1), 105 (27.1 %) were positive for skin Mf. The distribution of microfilaridemic positivity within the study communities were statistically significant (P = 0.001). The villages of Boalondo/Botatango had the highest level of endemicity (45.3 %). Of the 105 individuals positive for skin Mf, 73 were males and 32 were females. Males (31.9%) had significantly higher infection rates (P = 0.018) than females (20.1%). A total of 30.4% (95/312) of adults were positive for skin Mf compared with 13.2% (10/76) of children and the difference between these 2 groups was statistically significant (P = 0.007).
Table 1: Screened participants positive for O. volvulus Mf in the Bafia HD.
Variables
|
Study populations
|
Number examined
|
Number positive
|
Percentage (%)
|
P – value (x2test)
|
Sex
|
Male
|
229
|
73
|
31.9
|
0.018
|
Female
|
159
|
32
|
20.1
|
Total
|
388
|
105
|
|
Age- group
|
Children (5-19 years)
|
76
|
10
|
13.2
|
0.007
|
Adults (≥20 years)
|
312
|
95
|
30.4
|
Total
|
388
|
105
|
27.1
|
Communities
|
Balamba 1
|
28
|
7
|
25.0
|
0.001
|
Balamba 2
|
57
|
5
|
8.8
|
Biamo
|
68
|
23
|
33.8
|
Boalondo/Botatango
|
53
|
24
|
45.3
|
Lable/Nyamsong
|
88
|
12
|
13.6
|
Ngomo/Biatsota
|
94
|
34
|
36.2
|
Total
|
388
|
105
|
27.1
|
In the Melong HD (Table 2), the 599 participants comprised 467 adults, who were mainly farmers, and 132 children. The gender and age group distribution in each community were determined (S3 Table).
Following evaluation (Table 2), 70 (11.7%) participants were found positive for skin Mf. The villages of Ndoumbot/Ntangtom in the Melong HD had the highest proportion of infected persons (25.5%). Of the 70 individuals positive for skin Mf, 44 (15.5%) were males and 26 (8.2%) were females, and this difference was statistically significant (P = 0.005). Overall, 11.8% (55/467) of adults and 11.4% (15/132) of children were positive for Mf. Thus, the Bafia HD had significantly higher skin Mf proportions compared to the Melong HD (P = 0.002).
Table 2: Screened participants positive for O. volvulus Mf in the Melong HD.
Variables
|
Study populations
|
Number examined
|
Number positive
|
Percentage (%)
|
p – value (x2test)
|
Gender
|
Male
|
283
|
44
|
15.5
|
0.005
|
Female
|
316
|
26
|
8.2
|
Total
|
599
|
70
|
11.7
|
|
Age group
|
Children (5-19 years)
|
132
|
15
|
11.4
|
0.896
|
Adults (≥20 years) (
|
467
|
55
|
11.8
|
Total
|
599
|
70
|
11.7
|
|
Community
|
Mounko
|
60
|
4
|
6.7
|
0.002
|
Manjibo
|
51
|
12
|
23.5
|
Singa/Mbie/Barembeng2/Longze
|
104
|
10
|
9.6
|
Ndoumbot/Ntangtom
|
55
|
14
|
25.5
|
Nkoniakoniama/Nkonianke/Nkoniambot
|
87
|
7
|
8.0
|
Ndom-Bakem
|
79
|
10
|
12.7
|
Barembeng1
|
101
|
8
|
7.9
|
Mpaka
|
62
|
5
|
8.1
|
|
Total
|
599
|
70
|
11.7
|
|
From the 987 persons screened in the Bafia and Melong HDs, 175 microfilaridemic individuals were identified as potential candidates for the follow-up study. However, follow-up samples were only available from those who voluntarily consented.
In total 51 and 44 individuals could be recruited for the follow up study from the Bafia and Melong HD, respectively. The socio-demographic data from those individuals are shown in S4 and S5 Tables. Comparison analyses were made only for eligible participants (51, 51 and 48 at 1, 3 and 6 months post-treatment in the Bafia HD as well as 29, 44 and 38 at 1, 3 and 6 months after treatment in the Melong HD) from whom samples were collected at each of follow-up time points. A total of 150 skin snip samples from Bafia HD and 111 skin snip samples from Melong HD were collected for analysis (S6 table). None of the microfilaridemic participants were co-infected with M. perstans or L. loa as determined by thick blood films.
Susceptibility of O. volvulus microfilariae to ivermectin in two areas of contrasting MDA history following direct observed IVM treatment
As shown on Fig 3A, IVM led to a rapid drop in the proportion of microfilaridemic individuals after one month of treatment in both study sites. The proportion of participants detected with infection at the 1-month follow-up was significantly higher in the Bafia HD (21.6%) compared to the Melong HD (6.9%). The repopulation trend showed an increase in proportions of microfilaridemic individuals over time (D30 prevalence < D90 prevalence < D180 prevalence) by all techniques. However, the repopulation rate was faster in Bafia than Melong HD, presented by the larger surface area under the curves for geometric mean Mf density (Fig 3B) and Mf reduction rate (Fig 3C).
In detail, the pre-treatment geometric mean density for the 51 individuals recruited for follow-up in Bafia HD was 3.331 Mf/ss. At 1-month follow-up, the geometric mean Mf density per skin snip (GMMfD/ss) was 0.442 Mf/ss. When the 1-month GMMfD/ss was compared to the pre-treatment level, we observed 13.3% of the initial density (86.7% Mf reduction rate) in the geometric mean skin Mf (Fig 3C). IVM treatment elicited a statistically significant difference in the geometric mean density of skin microfilaridemia after 30 days (P < 0.001). At day 90 post treatment, all 51 follow-up participants were again examined and the GMMfD/ss at this time point was 0.379 Mf/ss. When compared to the pre-treatment level of 3.331 Mf/ss, we had 11.4% of the initial density left which resulted to 88.6% Mf reduction rate. After six months, 48 participants were examined for the fourth and last time of the study and the GMMfD/ss was 1.353 Mf/ss. The Mf reduction rate at 6 months was 59.4% with a skin microfilaria repopulation rate of 40.6%. The reduction in geometric mean density of skin microfilaridemia after 6 months was still significant (P = 0.01). The trend in geometric mean Mf reduction rate at different follow-up time points and repopulation of skin Mf shows a sharp decrease (86.7%) within 1 month of treatment, a further but slower decrease to 88.6% between one and three months and a rapid repopulation rate of 40.6% between the third and the sixth months after treatment (Fig 3B).
In the Melong HD pre-treatment geometric mean density for the 44 individuals recruited for follow-up was 2.318 Mf/ss. At 1-month follow-up, it decreased to 0.051 for the 29 participants examined giving a reduction rate of 97.8% (Fig 3C). Repopulation gradually sets in at the 3-month follow-up at a rate of 9.4%, but remained low even after 6 months (15%). The trend in geometric mean Mf reduction rate at different follow-up time points and repopulation of skin Mf (Fig 3B) shows a sharp decrease (97.8%) within 1 month of treatment, and a very slow repopulation rate from three to sixth months after treatment (Fig 3B and 3C).
Monitoring skin Mf rates after IVM treatment using microscopy, PCR and LAMP technology
In the Bafia HD, 51 eligible individuals were present for the follow-ups and a total of 150 skin snips were collected (Table 3).
Table 3: Proportion of microfilaridemia positivity detected by microscopy, PCR and LAMP at day 30, 60 and 180 post IVM treatment in Bafia HD
Time point after treatment (days)
|
Number of consented eligible participants
|
Number of samples collected
|
Proportion positive for microscopy n(%)
|
Proportion positive for real-time PCR n(%)
|
Proportion positive for O-150 LAMP n(%)
|
D30
|
51
|
51
|
11 (21.6%)
|
12 (23.5%)
|
15 (29.4%)
|
D90
|
51
|
51
|
12 (23.5%)
|
16 (31.4%)
|
19 (37.3%)
|
D180
|
51
|
48
|
22 (45.8%)
|
26 (54.2%)
|
31 (64.6%)
|
Total
|
|
150
|
45 (30%)
|
54 (36.0%)
|
65 (43.3%)
|
Significance
|
|
|
P = 0.014
|
P = 0.005
|
P = 0.001
|
n = number of cases detected by each method.
Proportion of microfilaridemia positivity detected by microscopy were 21.6% (11/51) after one month, 23.5% (12/51) after three months, and 45.8% (22/48) after six months. There was a statistically significant difference (P = 0.014) in the prevalence levels obtained at the different times of screening.
For the real-time PCR, a sample was considered positive when the actin signal (in duplicate) was above the threshold level. A representative set of samples from individuals are shown in S3 Fig. This method detected microfilaridemia proportion positivity of 23.5% (12/51), 31.4% (16/51) and 54.2% (26/48) at 1-, 3- and 6-months post treatment, respectively. There was a significant difference in the real-time PCR prevalence at the different time points (P = 0.005).
A colorimetric O-150 LAMP assay with a simple visual readout (S4 Fig) was also used to detect infection in DNA extracted from skin snips. Prior to initiation of the amplification reaction, samples were pink. After a 60 min incubation at 64 °C, samples turned yellow in the presence of O. volvulus DNA (S4A Fig). Samples remained pink in the absence of template, or if DNA from M. perstans or L. loa was present (S4B Fig), confirming the specificity of the LAMP assay. Colorimetric LAMP detected infection in 15/51 individuals (29.4%) at 1-month following chemotherapy, while 19/51 (37.3%) and 31/48 (64.6%) were detected after 3- and 6-months, respectively. A significant difference in the proportion of skin microfilaridemia positivity was also observed between the follow-up time points when using LAMP to detect infection (P = 0.001).
Despite the difference in targets for determining parasite infection, in Bafia HD, microscopy, real-time PCR and LAMP showed the same trend post treatment, namely a steady increase in the proportion of microfilaridermia positivity 1-, 3- and 6-months post treatment (Fig 4). The molecular assays were found to be more sensitive than microscopy, with highest levels of sensitivity obtained using LAMP. At each time point, the results obtained by comparing the different diagnostic methods were consistent (Microscopy < real-time PCR < LAMP) though not always significantly different (Fig 4). However, significant differences were seen when comparing two techniques (in pairs) at each time point (Fig 4).
In the Melong HD, 44 eligible recruited individuals were present for the follow-up study and a total of 111 skin snip samples (Table 4) were collected from them at 1 month (29), 3 months (44) and 6 months (38) post direct observed treatment, respectively.
Table 4: Proportion of microfilaridemia positivity detected by microscopy, PCR and LAMP at day 30, 60 and 180 post IVM treatment in Melong HD
Time point after treatment (days)
|
Number of consented eligible participants
|
Number of samples collected
|
Proportion positive for microscopy n(%)
|
Proportion positive for real-time PCR n(%)
|
Proportion positive for O-150 LAMP n(%)
|
D30
|
44
|
29
|
2 (6.9%)
|
4 (13.8%)
|
7 (24.1%)
|
D90
|
44
|
44
|
9 (20.5%)
|
7 (15.9%)
|
12 (27.3%)
|
D180
|
44
|
38
|
9 (23.7%)
|
9 (23.7%)
|
16 (42.1%)
|
Total
|
|
111
|
20 (18.0%)
|
20 (18.0%)
|
35 (31.5%)
|
Difference
|
|
|
P = 0.180
|
P = 0.520
|
P = 0. 215
|
n = number of cases detected by each method.
Microscopy detected proportions of Mf positivity of 6.9% (2/29) after one month, 20.5% (9/44) after three months and 23.7% (9/38) after six months. There was no significant difference in the positive proportions by microscopy at the different follow-up time points (P = 0.18), however the positivity increased with time from 1 to 6 months post treatment. Real-time PCR detected 4/29 (13.8%), 7/44 (15.9%) and 9/38 (23.7%) at 1-, 3- and 6-months post treatment, respectively. Colorimetric LAMP detected infections in 7 of the 29 individuals (24.1%), 12/44 (27.3%) and 16/38 (42.1%) after 1-, 3- and 6-months post treatment, respectively. No significant difference was observed in the positive proportions between the follow-up times regardless of the methods (Fig 5). However, after combining the samples collected at all points of monitoring (111), using the Chi square test, a significant difference (P = 0.02) was observed between the results from the three diagnostic methods. When comparing the performance of the techniques in pairs (e.g. Microscope Vs real-time PCR or LAMP Vs real-time PCR) significant differences in detecting infection where observed (Fig 5). Here too, LAMP assay detected the highest microfilaridemia proportions compared to the other methods at each time point (Fig 5).
An overview of the individual results of the three different methods and time points are shown in S6 table.
Performance characteristics of real-time PCR and LAMP assay using microscopy as the imperfect gold standard
The observed frequencies of various test result combinations were determined and entered into the data input template (S1 Table) of the Web-based application for each follow-up time point (D30, D90 and D180). The 3x positive Serial Number 1 (SN1) and 3x negative (SN8) have the highest numbers showing that PCR and LAMP confirm microscopy values. In addition, only positive in LAMP assay have the third highest observed frequency suggesting that the LAMP assay has a higher sensitivity compared to microscopy and qPCR.
Finally, there are no observed frequencies when the LAMP assay is negative whereas microscopy and qPCR are positive (SN2), further confirming the high sensitivity of LAMP.
The accuracy (sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV)) of real-time PCR and LAMP using microscopy as imperfect gold standard for skin biopsy samples collected from Bafia and Melong HDs are summarized in tables 5 and 6, respectively.
Table 5: Performance characteristics of the three tests in Bafia HD
Summary of sensitivity, specificity, PPV and NPV for microscopy, real-time PCR and LAMP at different time points following direct observed treatment of microfilaridermic individuals with ivermectin in Bafia health district.
METHODS
|
SENSITIVITY
|
SPECIFICITY
|
PPV
|
NPV
|
D30
|
|
|
|
|
MICROSCOPY
|
89.6 (61.6 - 99.9)
|
99.3 (93.0 - 100)
|
97.6 (76.9 - 100)
|
96.9 (86.2 - 100)
|
REAL-TIME PCR
|
89.6 (64.5 - 99.6)
|
96.8 (87.5 - 99.9)
|
89.6 (62.2 - 99.7)
|
96.8 (87.2 - 99.9)
|
LAMP ASSAY
|
97.6 (78.7 - 100)
|
92.1 (80.6 - 98.8)
|
79.0 (53.2 - 97.1)
|
99.2 (91.9 - 100)
|
D90
|
|
|
|
|
MICROSCOPY
|
71.4 (45.8 - 91.8)
|
96.9 (88.0 - 99.8)
|
90.6 (67.0 - 99.5)
|
88.8 (74.8 - 97.3)
|
REAL-TIME PCR
|
97.5 (75.8 - 100)
|
97.0 (86.3 - 100)
|
93.2 (69.3 - 99.9)
|
98.9 (87.8 - 100)
|
LAMP ASSAY
|
97.5 (79.0 - 100)
|
88.7 (75.0 - 97.6)
|
78.6 (54.5 - 95.7)
|
98.8 (89.1 - 100)
|
D180
|
|
|
|
|
MICROSCOPY
|
80.6 (62.9 - 93.1)
|
94.8 (81.3 - 99.6)
|
94.9 (80.6 - 99.6)
|
80.8 (61.5 - 93.4)
|
REAL-TIME PCR
|
98.9 (87.3 - 100)
|
98.5 (84.6 - 100)
|
98.8 (86.4 - 100)
|
98.7 (83.4 - 100)
|
LAMP ASSAY
|
99.0 (90.1 - 100)
|
76.6 (56.3 - 91.6)
|
83.4 (66.9 - 94.3)
|
98.6 (85.2 - 100)
|
D30+D90+D180
|
|
|
|
|
MICROSCOPY
|
80.7 (68.4 - 90.2)
|
97.4 (93.1 - 99.5)
|
94.5 (85.1 - 98.9)
|
90.3 (83.1 - 95.2)
|
REAL-TIME PCR
|
98.1 (90.0 - 100)
|
97.6 (92.7 - 99.8)
|
95.7 (87.1 - 99.6)
|
99.0 (94.2 - 100)
|
LAMP ASSAY
|
99.4 (93.4 - 100)
|
87.1 (79.4 - 93.1)
|
80.7 (69.6 - 89.9)
|
99.6 (95.8 - 100)
|
Table 6: Performance characteristics of the three tests in Melong HD.
Summary of sensitivity, specificity, PPV and NPV for Microscopy, real-time PCR and LAMP at different time points following direct observed treatment of microfilaridermic individuals with ivermectin in Melong Health District
METHODS
|
SENSITIVITY
|
SPECIFICITY
|
PPV
|
NPV
|
D30
|
|
|
|
|
MICROSCOPY
|
45.4 (10.1 - 91.2)
|
99.1 (89.7 - 100)
|
89.6 (29.3 - 100)
|
91.1 (71.0 - 99.2)
|
REAL-TIME PCR
|
87.6 (32.2 - 100)
|
98.7 (87.1 - 100)
|
92.2 (36.2 - 100)
|
98.0 (78.3 - 100)
|
LAMP ASSAY
|
94.2 (52.0 - 100)
|
89.3 (71.5 - 99.7)
|
60.5 (18.4 - 99.3)
|
98.9 (87.4 - 100)
|
D90
|
|
|
|
|
MICROSCOPY
|
90.6 (48.3 - 100)
|
94.7 (82.3 - 100)
|
78.9 (35.2 - 99.9)
|
98.0 (82.0 - 100)
|
REAL-TIME PCR
|
38.0 (10.9 - 80.6)
|
99.4 (92.9 - 100)
|
93.1 (43.4 - 100)
|
87.8 (69.8 - 97.9)
|
LAMP ASSAY
|
92.1 (55.3 - 100)
|
89.8 (75.9 - 99.7)
|
66.1 (28.1 - 99.2)
|
98.1 (84.9 - 100)
|
D180
|
|
|
|
|
MICROSCOPY
|
88.9 (58.1 - 99.9)
|
96.1 (84.9 - 99.8)
|
87.5 (59.2 - 99.4)
|
96.5 (84.5 - 100)
|
REAL-TIME PCR
|
96.7 (69.9 - 100)
|
98.8 (89.2 - 100)
|
96.2 (66.7 - 100)
|
99.0 (89.4 - 100)
|
LAMP ASSAY
|
97.1 (74.0 - 100)
|
75.3 (58.7 - 88.5)
|
55.0 (30.6 - 78.2)
|
98.9 (87.7 - 100)
|
D30+D90+D180
|
|
|
|
|
MICROSCOPY
|
80.2 (57.2 - 95.0)
|
95.9 (89.4 - 99.2)
|
81.5 (56.6 - 96.4)
|
95.7 (87.9 - 99.0)
|
REAL-TIME PCR
|
78.3 (53.3 - 97.9)
|
99.7 (96.5 - 100)
|
98.2 (82.1 - 100)
|
95.4 (87.4 - 99.7)
|
LAMP ASSAY
|
98.5 (84.3 - 100)
|
84.7 (75.6 - 92.3)
|
58.7 (39.1 - 79.4)
|
99.6 (95.4 - 100)
|
The sensitivity of LAMP was generally higher than that of real-time PCR in detecting O. volvulus infection following treatment with a microfilaricide (IVM) in both study sites (Tables 5 and 6) at different time points. Bafia HD recorded sensitivities of 99.4% for LAMP, 98.1% for real-time PCR and 80.7% for microscopy while Melong HD had 98.5% for LAMP assay, 78.3% for real-time PCR and 80.2% for microscopy. With the exception at 3 months follow up in Melong HD, molecular methods had higher and comparable sensitivities. In both study sites, microscopy and real-time PCR techniques showed very high and comparable specificity values (always >94 %) while LAMP is relatively lower at all the follow-up time points (ranging between 75 – 92 %). With exception at 1-month follow-up in Bafia HD, the real-time PCR assay had the highest PPV of over 92% at the different time points. The LAMP had relatively lower PPV compared to real-time PCR and microscopy. In terms of negative predictive value, LAMP had the highest values (ranging between 98.1 – 99.6 %) while the NPV of microscopy and real-time PCR were relatively lower and comparable.