Microclimate variables of the ambient environment deliver the actual estimates of the extrinsic incubation period of Plasmodium vivax and Plasmodium falciparum: a study from a malaria-endemic urban setting, Chennai in India

The catchment area of Besant Nagar clinic (13.0002°N, 80.2668°E) was selected for the study based on the malaria prevalence during the period, 2006–2012 obtained from the Regional Office for Health and Family Welfare (ROH and FW), Besant Nagar, Chennai. Appropriate sites were selected for year-long environmental monitoring to cover all the possible microclimatic regimes of the various structure types in indoor and outdoor environments of the study area (Fig. 1). The map of the study area with hobo locations was generated with the help of Google Earth Pro v7.1. These included the common household roof structure types like thatched, tiled, asbestos and concrete material besides, other outdoor structures like overhead tanks (OHTs) and wells, the potential breeding habitats of An. stephensi and the vegetation. The study was conducted from November 2012 to October 2013 in Besant Nagar, a residential area with slums adjacent to the seashore in the south-eastern part of Chennai; distinctly characterized by its meso-endemic perennial transmission of malaria, predominantly P. vivax.

Microclimatic temperature and relative humidity (RH) of the resting environments that adult An. stephensi are presumed to experience were recorded using 42 temperature/RH data loggers (Onset HOBO U10-003) on an hourly basis. They were equally distributed among seven structure types with six replicates including indoor and outdoor environments of human dwellings with varied roof structures and three selected outdoor resting sites of OHTs, wells and vegetation. Back-up for each structure type was also included in case of any missing or malfunction readings. Prior to the placement of data loggers, informed consent was obtained from the household members for keeping them year-long at the same location without any disturbance. Hobos were labelled with numbers (1–42) and distributed in the field site with details of each location documented for quick reference/identification. The coordinates of each logger were recorded using a Global Positioning System (GPS: Garmin-version 2.40). Hobos were either hung on the nails available on the walls or were kept over an open, horizontal flat surface, one to two feet down, from the roof for indoor and similarly in outdoor structures. Care was taken to place the data loggers at locations away from sources of heat and moisture in houses, such as laundry, showers and cooking area, which would provide inaccurate readings of a particular area [21]. In wells, the HOBOs were hung above the normal and expected increase of water level during monsoon period after confirming it with the house owners. Similarly, in the OHTs, the HOBOs were placed above the water level to avoid submerging of the HOBOs. Routine monthly visits were made during the study period in the forenoons and the data from each logger was downloaded on to a laptop with a software, Hoboware Lite (Ver. 3.2.1). While downloading, the data was checked for any errors; an abnormal or large number of missing readings and also the functioning besides, the battery conditions of the HOBO data loggers. If anything was found abnormal, the respective data was excluded and the HOBO was reset, ran for a short time and checked for readings. If the readings were recorded properly, the device was re-launched and placed at the same location. Malfunctioning or missing loggers were replaced with new backups. Further, after obtaining a year-long dataset, data cleaning was performed to remove any missing, abnormal values. The excluded data from the malfunctioned hobos were compensated by replacing them with the corresponding data points from the backup hobos placed in the respective structure types.

The monthly malaria prevalence for the study period at the study site was obtained from Regional Office for Health and Family Welfare, Government of India at Besant Nagar, Chennai. The man-hour density (MHD) for the study period was obtained from a parallel study of the fortnightly collections of the adult female An. stephensi mosquitoes in cattle sheds during the dusk period. The monthly MHD of An. stephensi was calculated by dividing the total number of female mosquitoes collected by total time spent for a particular month for one hour period i.e., (Total female An. stephensi collected/Total time spent) × 60 min [22]. The monthly rainfall data was obtained from Regional Meteorological Centre at Nungambakkam, Chennai, India.

The downloaded data points were arranged on the basis of individual hobos, day and month wise and also structure types. The maximum, minimum and mean values of the temperature and RH of structure types were calculated based on hour, day and logger wise besides, on monthly basis. Daily temperature range (DTR) is considered as the difference between the highest and lowest values of temperatures, recorded during a day while daily relative humidity range (DRHR) is the difference between highest and lowest values of relative humidity recorded during a day. The monthly mean DTR was calculated as the mean of all DTR values of the corresponding month. The readings logged between 6.00 and 17.59 h were considered as day-time/diurnal readings when anophelines, in general, are considered to be primarily resting and those recorded during 18.00 and 5.59 h were night-time/nocturnal readings when they are in an active phase [23]. Hour, month, indoor–outdoor, structure type and season-wise differences in temperature, RH, DTR and DRHR as well as their diurnal and nocturnal variations were analysed statistically (Independent t-test and ANOVA). The environmental monitoring data was processed by IBM SPSS version 21. Further, extrinsic incubation period (EIP) was derived from the longitudinal temperature readings observed in the study site. EIP is defined as the interval between the acquisition of parasite by the vector and the vector’s ability to transmit the parasite to other susceptible vertebrate hosts. It is calculated as the reciprocal of parasite development rate (PDR). PDR of P. vivax and P. falciparum was estimated using the following equation [9].

According to the thermodynamic model, 34.4 °C (P. vivax) and 35 °C (P. falciparum) were considered as the critical maximum temperature wherein parasite developments were assumed to be blocked if mean temperatures exceeded the respective CT_max [9]. Since the temperature data points above those maximum limits (43 data points out of 4015 in the present dataset) were assumed to generate invalid estimates of EIP, they were excluded from the EIP estimation analysis to make the data more reliable and accurate. Further, the relationships among mean temperature, RH, DTR, DRHR, rainfall, MHD, EIP and monthly malaria prevalence were analysed using Pearson correlation test.

The microclimate (temperature and RH) profile observed during the year-long study revealed variations among the structure types in indoor and outdoor environments. The recorded ambient temperature and RH represent the actual microclimate conditions experienced in these structure types with a wide range of DTR, DRHR and also variations in diurnal and nocturnal temperature in all the structure types found indoors and outdoors.

The average hourly, monthly temperature profile and RH recorded by the data loggers indicated that indoor temperatures were warmer when compared to outdoors with less humidity (Fig. 2). Among the structure types of roof materials (households), the indoor mean temperature ranged from 27.63 ± 0.62 °C in thatched (Jan. ‘13) to 33.66 ± 1.05 °C in tiled structure (May ‘13). In the outdoor structures of households, the corresponding mean temperature ranged from 26.64 ± 0.44 °C in thatched (Jan. ‘13) to 33.22 ± 0.93 °C in tiled (May ‘13). Thatched (indoor and outdoor) structure was cooler with a low-temperature profile in contrast to tiled structure, with a warmer ambient environment of high temperature. In the case of other outdoor structures (OHTs, wells and vegetation), the lowest temperature of 26.72 ± 0.8 °C (Jan. ‘13) was observed in vegetation and the highest of 34.15 ± 2.29 °C (May ‘13) in the overhead tank (Table 1). The DTR observed a wide temperature range in other outdoor structures ranging from 0.89 °C (Dec. ‘12) in wells to 14.62 °C (Mar. ‘13) in OHTs (Table 1). In indoors, DTR ranged from 1.93 °C (Sept. ‘13) in concrete to 7.07 °C (June ‘13) in thatched structure whereas, in outdoors, it ranged from 2.3 °C (Dec. ‘12) in concrete to 12.01 °C (Feb. ‘13) in asbestos structure.

However, RH observed a different picture in indoor household structure types ranging from 57.66 ± 8.01% (June ‘13) in the concrete structure to 76.77 ± 2.69% (Apr. ‘13) in thatched structures. Similarly, RH in outdoor household structure types ranged from 62.07 ± 8.05% (June ‘13) in concrete to 81.45 ± 8.6% (Oct. ‘13) in thatched structures. Among all structure types, thatched structures were more humid, both indoors and outdoors whereas, concrete structures experienced the lowest humidity. In other outdoor structures, RH was lowest in vegetation (67.77 ± 11%) during June ‘13 and highest in well (96.14 ± 3.44%) during Oct. ’13 (Table 1). The mean temperature was warmer and obvious during the early part of the day (morning) and later remained almost constant (evening and night) in indoors (Fig. 2a, b). However, outdoors were more humid than indoors and varied drastically during the 24 h period (Fig. 2c) as well as throughout the study period (Fig. 2d).

Similarly, month wise observation of microclimate profile (temperature and humidity) experienced by indoor and outdoor environments among various structure types are detailed in Table 1 and Fig. 3. In indoors, thatched structure observed the minimum temperature profile throughout the study period whereas, tiled exhibited the highest temperature in May ‘13 with fluctuations during the initial (Nov. ‘12 to Mar. ‘13) and later part (June ‘13 to Oct. ‘13). In outdoors, the lowest temperature was recorded for thatched (Jan. ‘13) and the maximum temperature was consistent for OHTs (Fig. 3a). Mean DTR had maximum variations by and large for thatched and the least for concrete in indoor structures. Similarly, the mean DTR had the highest variations in OHTs and the lowest in wells among all other outdoor structures (Fig. 3b). The mean RH recorded, indicated a uniform pattern in indoors for all structure types, however in outdoors, wells recorded the maximum (Fig. 3c). The mean DRHR indicated maximum for indoor thatched structure whereas, in outdoors, tiled structure recorded the maximum (Fig. 3d).

When the mean values of each of the variables (temperature, RH, DTR, DRHR) were compared across different combinations like (i) all indoors versus all outdoors (ii) indoors of human dwellings/households versus the corresponding outdoor environments and (iii) human dwellings in general versus outdoor resting habitats alone, it was observed that in all the three combinations, mean temperature of indoors and human dwellings in general (indoor and outdoor) were observed to be significantly higher compared to outdoors (Table 2). However, RH and DTR of all household structure types (indoor and outdoor) were low whereas, all outdoor structure types were significantly higher among the all-different combinations (p < 0.001).

Similarly, when the structure wise difference in the indoor and outdoor environment was analysed for temperature, RH, DTR and DRHR (Table 3), it was observed that all the variables were significantly different in the indoor and outdoor environment and also between indoors and outdoors of the different structure types (p < 0.001). When the mean temperatures of OHTs were compared with the temperature of all indoors in general and also with each indoor structure type (tiled, asbestos, concrete, thatched), it showed a significant difference (p < 0.001). Further, when the mean temperature of wells when compared with the temperatures of all indoors together and with each indoor structure type, it showed a significant difference (p < 0.001) in the temperature profile.

It was observed that, across the months, diurnal and nocturnal temperature besides, RH had a significant difference (p < 0.001). The day temperatures were significantly higher, except concrete indoors, compared to nights (p < 0.001) across various structure types in both indoor and outdoor environments (see Additional file 1), while RH was almost the same without much difference (see Additional file 2). Further during the day-time, indoors and outdoors exhibited similar temperature pattern, whereas, during the night, the outdoor temperature was observed to be less unlike indoors, where it remained almost the same. The indoor/outdoor and structure type wise variations in the examined variables were found to be significantly different across seasons (p < 0.001). Outdoors remained more humid during day and night (see Additional file 2). When the diurnal and nocturnal variations in temperature and RH was compared season wise, it was found that except for temperature during monsoon period (Day-time = 30.29 °C; 95% CI 28.74–31.85 °C; nocturnal = 28.48 °C; 95% CI 26.84–30.12 °C), there was no significant difference in any of the variables.

It was observed that, during all seasons, the diurnal temperature was significantly different from nocturnal temperature, i.e., winter (mean diurnal temperature = 28.70 °C, 95% CI 28.67–28.72 °C; mean nocturnal temperature = 27.47 °C, 95% CI 27.45–27.49 °C); summer (mean diurnal temperature = 31.53 °C, 95% CI 31.50–31.56 °C; mean nocturnal temperature = 31.20 °C, 95% CI 31.17–31.22 °C); pre monsoon (mean diurnal temperature = 31.54 °C, 95% CI 31.50–31.57 °C; mean nocturnal temperature = 29.69 °C, 95% CI 29.67–29.71 °C); monsoon (mean diurnal temperature = 30.30 °C, 95% CI 30.27–30.33 °C; mean nocturnal temperature = 28.49 °C, 95% CI 28.47–28.51 °C). Similarly, during all seasons, the diurnal RH was significantly different from nocturnal RH, i.e., winter [mean diurnal RH = 75.80% (95% CI 75.60–75.89%); mean nocturnal RH = 76.33% (95% CI 76.24–76.42%)]; summer [mean diurnal RH = 76.31% (95% CI 76.21–76.41%); mean nocturnal RH = 76.59% (95% CI 76.49–76.69%)]; pre monsoon [mean diurnal RH = 73.13% (95% CI 72.99–73.27%); mean nocturnal RH = 76.73% (95% CI 76.61–76.84%)]; monsoon [mean diurnal RH = 78.32% (95% CI 78.21–78.43%); mean nocturnal RH = 80.42% (95% CI 80.33–80.51%)].

The peak prevalence of P. vivax and P. falciparum illustrates a different picture and it is quite contradictory to the man-hour density observed in the study site (Fig. 4). The estimated EIPs of various structure types in human dwellings (Fig. 5) exhibited a wide range in EIPs across indoors (P. vivax—7.97–14.08; P. falciparum—9.11–12.52) as well as outdoors (P. vivax—7.97–11.61; P. falciparum—9.11–11.30). EIPs of indoor structure types of human dwellings (except thatched) were observed to be more when compared to outdoors. However, there was no significant difference in EIP when comparisons were made among structure types of human dwellings (indoor and outdoor) and also with structure types of human dwelling outdoors against all outdoor structures (Table 4). Among the other outdoor structures, EIP of P. vivax was estimated to be 7.97–24.27 and for P. falciparum, it was 9.11–15.26 in the overhead tank (Table 4). OHT exhibited maximum temperature variations and estimates of EIP for P. vivax and P. falciparum, while wells recorded the lowest estimates of temperatures for EIP in the case of P. falciparum. Further, the mean EIP values of both malaria parasites derived from all indoors were compared against the EIP values of potential breeding habitats like OHTs and wells. It was observed that the estimated EIP of P. vivax derived from indoors (Mean EIP = 8.667) were significantly different (p = 0.040), from the EIPs derived from OHTs (Mean EIP = 10.24). However, there was no significant difference between indoors and wells. Further, when EIP values of individual indoors were compared against wells and OHTs, it was observed that EIP of P. vivax derived from thatched indoors (Mean EIP = 9.39) were significantly different (p = 0.048), from the EIPs derived from OHTs (Mean EIP = 10.24).

The trend of temperature, RH, rainfall, DTR, MHD and average malaria prevalence (2006–2012) during the study period has been represented in Additional file 3. It was observed that, as mean temperatures increases, DTR showed a significant positive correlation (r = 0.422, p < 0.001) in general and also during different months of a year (r = 0.694, p = 0.01). However, when the data was analysed month wise, RH showed significant negative correlation with MHD (r = − 0.661, p = 0.019) and DTR (r = − 0.661, p = 0.01). Also, DTR showed significant positive correlation with DRHR (r = 0.618, p = 0.032). Further, temperature showed significant positive correlation with DTR (r = 0.694, p = 0.012) and DRHR (r = 0.755, p = 0.005).

The various parameters of temperature, relative humidity, DTR and DRHR had a high significant difference (p < 0.001) within each structure type except few which had a less significant difference with p < 0.05 whereas some were non-significant, which are depicted in Fig. 6. Only OHT had a significant difference (p < 0.05) with other structure types when EIP of P. vivax and P. falciparum was compared with different structure types.

Mean temperature differed significantly in each structure type except indoors of asbestos, concrete and tiled. However, mean RH differed significantly in all structure types except indoors of thatched and tiled, outdoors of asbestos, concrete and tiled. Further mean DTR differed significantly in each structure type except concrete outdoor and tiled indoor. While DRHR did not differ significantly in indoors of asbestos and thatched, outdoors of concrete, thatched, tiled and OHT. EIP (P. vivax) did not differ significantly in different structure types except OHT which differed significantly with other structure types except tiled indoor. EIP (P. falciparum) also did not show any significant difference within different structure types except OHT which differed with indoors of concrete and thatched, outdoors of asbestos and concrete (Fig. 6).