Physical integrity of three Long-Lasting Insecticidal Net products 30 months after a mass-distribution campaign in Benin

Abstract Background The effectiveness of long-lasting insecticidal nets (LLINs) mainly relies on their physical integrity, an important indicator of the LLINs durability in households. The present study aims to assess this physical integrity for three widely used LLIN products: DawaPlus®2.0, PermaNet®2.0 (polyester) and DuraNet® (polyethylene), distributed in Benin during the mass campaign of 2014. Methods The study was conducted from October 2014 to June 2017 in three districts of Benin: Tori-Bossito, Toffo, and Ouesse. Nine hundred LLINs (a cohort of 300 LLINs per product) in use and found hanging on sleeping materials were selected and tagged. Every six months, the LLIN attrition and their physical condition was monitored. The holes were counted by category allowed to calculate the proportionate hole index (pHI) and classify the LLINs into "good ", "serviceable" or "torn" categories. Factors associated with loss of integrity have also been identified following World Health Organization Guidelines. Results After 30 months of use, 55.9% (n= 503; 95% CI: 52.6-59.2) of the LLINs were lost. Attritions due to physical damage were similar between the LLIN products and were respectively 28.3% for PermaNet® 2.0, 29.7% for DawaPlus® 2.0, and 31% for DuraNet® (p˃0.05). The mean pHI was significantly higher for DuraNet® LLINs (pHI= 1431) than DawaPlus® 2.0 (pHI =366) and PermaNet® 2.0 (pHI =321) (<0.001). A significant difference was also observed between the proportion of LLINs in “torn” condition and included 8.7% of PermaNet® 2.0, 14.3% of DawaPlus® 2.0, and 34.1% of DuraNet® (p <0.001). The use of LLINs every night, the frequency of washing and other factors were significantly associated with the physical damage of the LLINs (p <0.001). Conclusion These results suggest that the physical barrier conferred by the LLIN can be significantly affected during the

choice of LLIN product to be distributed. Keywords: LLINs, physical integrity, pHI, malaria, Benin.

Background
In sub-Saharan Africa, the proportion of people with access to long-lasting insecticidal nets (LLINs) has increased from less than 2% in 2000 to 67% in 2015 [1]. To improve this performance observed in many countries in Africa and achieve the universal coverage goal, more than 4 million Olyset®Net LLIN were freely distributed to the population in 2011 in Benin during the national mass-distribution campaign [2]. The LLINs distributed act on adult mosquito vectors and aim to reduce or limit human-vector contact to prevent malaria [3]. After 2 years, 56% of them were removed from their initial households, for different reasons [4]. In addition, rapid physical deterioration of the LLINs was observed 24 months after the distribution, with 90% found with holes [4,5].
The level of rapid damage of these LLINs is a concern. A similar finding was observed In Uganda with 45-78% of the LLINs damaged after one year of use under operational conditions [6]. About 40% of LLINs were also damaged after 2 to 3 years of use in another study in Laos [7]. Although extrinsic factors related to human practices (washing of nets, type of bedding) and the immediate environment are responsible for the rapid deterioration observed on LLINs, the quality of their physical barrier could also be incriminated [8,9].
The Olyset®Net polyethylene LLINs distributed in 2011 were replaced in 2014 by three other types of polyester (DawaPlus®2.0, PermaNet®2.0) and polyethylene (DuraNet®) [10] LLINs. This offers an exceptional opportunity to simultaneously compare, for the first time in Benin, the performance of these LLINs in large-scale household use condition. The LLINs distributed are industrially pretreated using specific processes that allow them to be effective after at least 20 washes and to maintain their bioefficacy for 3 to 5 years in normal use [11].
The performance of the LLINs distributed is mainly based on three indicators: the attrition representing the proportion of LLINs lost; the physical integrity, an estimate of the number and size of holes the LLINs; and the bio-efficacy, a measurement of the insecticidal effect [9]. The term "hole" is used to describe all types of damage: rips burn holes, rodent damage, torn corners and stitching defects on LLINs [12]. The attrition and physical integrity of the LLINs, which represent the target indicators of this study, are critical indicators since they represent the major limiting factor in LLIN durability and could lead to a poor protection of the sleeper [4]. the opportunity to compare them simultaneously after a mass distribution campaign is are given the predominance of a single type of LLIN in mass distribution campaigns [14,15].
Several malaria control programs are therefore working with the assumption that the physical integrity of the LLIN varies little between the types of products offered and thus opts for the choice of the lowest bidder [15]. This longitudinal study undertaken by the Center for Entomological Research in Cotonou (CREC) reports the comparative results of the physical integrity of three types of LLINs under community use conditions. It aims to strengthen the capacity of Malaria Programs to select LLINs to be distributed based on their local performance.

Study sites
Following the LLINs distribution, a monitoring study on their physical integrity was initiated in three districts, Tori-Bossito, Toffo and Ouesse (Table 1). These three districts were randomly selected from a list of districts coded according to the type of LLINs received. This selection was made online using the Random Number Generator [16]. Two of the three districts selected are located in the Atlantic Department: Tori-Bossito in the South and Toffo in the North. The third district (Ouesse) is located in the department of Collines ( Figure 1). In each of the three districts, two locations (one rural and one urban) were selected from their location list by applying the same procedure (list of districts, numeric code, random number generator) for the selection of the districts previously mentioned (Table 1).

Study design
The study consists of longitudinal monitoring of the LLINs every 6 months and was

Household selection and tagging the LLINs (T0)
A total of 900 LLINs were randomly selected and marked on the basis of a cohort of 300 LLINs per net brand according to WHO guidelines [17] just after the mass-distribution campaign of October 2014. The consent to participate in the monitoring surveys was obtained from the head of each household or his spouse. Only one LLIN was selected per consenting household to be sure capture how variation in household practices and environment could affect the LLINs durability. The selection of households at each site took into account the household list of all villages to ensure representative sampling [9].
The assessment teams identified a LLIN distributed in 2014 in each selected household and verified whether the LLIN was hanged-up and its physical integrity was not yet compromised. Each selected LLIN has been tagged with a unique identification number to ensure proper identification during subsequent visits. During the monitoring of the integrity of the selected LLINs, the teams entered the homes where the LLINs were installed in accordance with World Health Organization (WHO) recommendations [18].

Monitoring survey and questionnaire
A model questionnaire developed by WHO [18] has been reviewed and adapted to the needs of the study. This questionnaire was used to identify household characteristics, sleeping material, level of education, etc. The questionnaire has been programmed on Samsung Galaxy 10.1 tablets with ODK Collect 1.2.2. Tablets were directly connected to the internet via a SIM card. They were used for the interviews of the head of households or his spouse but also for collecting data on the integrity of the LLINs (hole size and number of holes). These data were directly sent to a cloud server used to store the data which were retrieved later for data cleaning and analysis.

LLIN attrition
LLIN attrition represent the proportion of LLINs initially enrolled in the study and which have been lost due to different reasons. Attrition has been categorized by the main reasons why a LLIN is no longer used (attrition 1: destroyed, so torn and worn; attrition 2: stolen, given away, moved; and attrition 3: used for other purposes) [18].

Evaluation of the physical integrity of the LLINs
LLINs were examined visually, without being removed from where they were hung in the targeted houses. The total number of holes observed was counted and classified according to four general size categories: -Size 1: hole smaller than one inch (0.5-2.0 cm), -Size 2: hole larger than the thumb but smaller than the fist (2-10 cm), -Size 3: hole larger than the fist but smaller than the head (10-25 cm), -Size 4: hole larger than the head (> 25cm).
Descriptive statistics were used to compare pHI values at each assessment site. Based on the pHI value, LLINs were assigned to one of three categories [19]:

Statistical analysis
Data processing is done with R version 3.5.1 software (R Core Team, Vienna, Austria 2018) and Excel. The calculation of the proportions of the type of holes as well as their comparison between two locations were carried out using the chi 2 test. The binomial test was used to estimate the proportions and their confidence intervals. The Shapiro-Wilk test was used to assess the normality of the data. The Kruskal-Wallis test was used to compare the median pHI but also the performance of the LLIN products by category (good, usable, to replace and lost) for each period (6, 12, 18, 24 and 30 months). The test of equality of proportions was used to determine the existence of equality or not between the proportions of the different locations. The threshold of significance of the tests was 5%.

Assessment of holes
At the 6th month assessment visit, the LLINs with holes represented 29% (95% CI: 23. In both rural and urban settings, there was no significant difference between the different hole sizes for the three types of LLINs after 6, 18 and 24 months of use. However, a significant difference was observed for the size 3 holes for the PN 2.0 LLINs at 24 months and the size 4 holes for the Dura LLINs at 30 months. However, the hole proportions per size category varied from one location to another, among LLIN products, and over time ( Table 3). Table 4 showed that the pHI gradually increased from one period to another for the three types of LLINs and were: from 39.8 (at 6 months) to 366 (at 30 months) for Dawa 2.0, 63.6 (at 6 months) to 321 (at 30 months) for PN 2.0, and 176 (at 6 months) to 1431 (at 30 months) for Dura ( Figure 4). No significant difference in the pHI was observed from the 6 to 18 months for the three LLINs. A significant difference was noted from at the 24 and 30 months (p <0.001) when the highest pHI was noted.

Comparison of the proportionate hole index between LLIN products
The data showed that the means pHI of Dawa 2.0, Dura and PN 2.0 during the 12 and 18 months were similar (p> 0.05) (
At LLIN brand level, 39.1% (95% CI: 27.6-51.6) of PN 2.0 were "in good" condition, 26.1% (95% CI: 16.3-38.1) were "serviceable" and 8.7% (95% CI: 3.3-17.9) were in "torn" after 30 months of follow-up. In the same period 28.6% (95% CI: 18.4-40.6), 27.1% (95% CI: 17.2-39.1) and 14.3% (95% CI: 7.1-24.7) of the Dawa 2.0 were respectively in "good" condition, "serviceable" and "torn" conditions. For the Dura LLINs, the proportions of the LLINs in "good" condition, "serviceable" and "torn" conditions were respectively 15.9% (95% CI: 6.6-30.1), 34.1% (95% CI: 20.5-49.9) and 34.1% (95% CI: 20.5-49.9) ( Table 6). Table 7 shows that most of the recorded factors were significantly associate to the physical damages of the LLINs (p <0.05). LLINs often used are more holed than those used every night for DawaPlus 2.0 without statistical difference (p> 0.05) unlike Dura and PN 2.0 (p <0.05). Households where the sleeping beds are cut bamboo or palm branch, with poor maintenance of the LLINs (dirty), having the kitchen indoor close to the LLINs and which have a washing frequency more than 6 times have their LLIN with more holes than households where the mats were used as sleeping material, where the LLINs are well maintained (clean) and where the kitchen were outside with a washing frequency less than 6 times. There was a correlation between these factors and loss of the physical integrity of the LLINs (p <0.001).

Discussion
The key LLIN component preventing man-vector contact is the physical integrity [5].
Therefore, the loss of physical integrity represents a major threat for the sleeper and need strong attention. In this study, the loss of integrity observed after the mass distribution campaign in 2014 in Benin was so large that it could have compromised the effectiveness of the LLINs in preventing malaria. The proportionate hole index (pHI) [19] provides a standardized approach to describe changes in the tissue integrity. Our results showed that during the 30 months follow-ups, there was a significance difference among the three LLIN products assessed in term of pHI which exponentially increase over time. After 30 months of use, the pHI of Dura (polyethylene) (pHI= 807.21) were higher than those of Dawa 2.0 (pHI= 185.75) and PN 2.0 (polyester) (pHI= 183.15) suggesting that Dura LLINs were more likely to lose their physical integrity.
The categorization of these pHI according to the WHO standards [20] as being in "good condition", "damaged but usable" or "to replace" showed us that after 30 months of use, 29.5 % of the LLINs (all brands) were in good condition. This result was higher than those observed in Madagascar in 2012 [21] where the proportion of LLINs to be replaced after 30 months was 16 LLINs. This finding confirms the results observed in Mozambique (1-year follow-up) [22] but was not consistent with the other findings which reported a low physical damage rate of a polythelene-based LLIN (Olyset net) after 7 years of use [23] suggesting that the loss of physical integrity may vary depending of the local conditions. The frequency of LLIN use (often or every night), the type of sleeping material (mat or bed), the appearance of the LLIN (clean or dirty), the location of the kitchen (indoor or outdoor) and the frequency washing were correlated with loss of tissue integrity. These different practices in the use and maintenance of LLINs vary from one community to another and probably affect the status of different types of LLINs [24,25,26,27,28,29].
It is therefore important to strengthen the awareness of the best practices for the maintenance of LLINs. In Kenya, a recent study found that people washed their LLINs more often than recommended and that this practice was associated with poor physical integrity of the LLINs [9]. It was also revealed that light-colored mosquito nets were more likely to be washed than dark-colored mosquito nets. The polyethylene (Dura) based LLINs in this evaluation were light blue. The majority of LLINs with high pHI values were dirty and often used. Thus, it is possible that these factors also contributed to the observed high frequency of washing that resulted in relatively high pHI of the LLINs washed more than 6 times and that significantly affected the physical integrity (p <0.001). WHO recommends that LLINs should be monitored from 6 months [30] after distribution to understand the reasons for attrition and wear in the first year of use. Although this was the case, the study was not able to compare how the extrinsic factors affect the physical integrity of each LLIN product and represents an important limit for the study.

Limitations
The prospective and longitudinal nature of this study may have led participants to retain LLINs longer than expected if they were not enrolled and monitored. As a result, attrition rates due to LLIN disposal would have been higher and the loss of physical integrity observed may be worse than that normally tolerated before net disposal. A large number of LLINs were no longer available for follow-up for many reasons and could have skewed the results. The LLINs that were monitored were in three different regions. The differences observed between the types of LLINs, such as use and maintenance practices, probably reflect the differences in circumstances and regional customs. The results regarding LLIN repair were not taken consider as the LLINs in this study were poorly repaired as previously reported in many other studies [22,31]. This study did not also take into account LLIN repair and their effect on physical durability as this practice was not common in the study communities and very few data were recorded. The insecticidal effectiveness of the LLINs and the prevalence of specific damage mechanisms were not also estimate representing an important limit for the study.

Conclusion
The physical integrity of the three LLIN products has been relatively low with a higher damage observed in the polyethylene-base LLINs: DuraNet. LLINs are often used to sleep on beds, often too dirty and washed, but rarely repaired by their owners. Awareness and emphasis on washing frequency, as well as the usefulness of LLINs even with holes, could         Variation in pHI of the three types of LLINs over time