Advances in mosquito repellents: effectiveness of citronellal derivatives in laboratory and field trials

Abstract

BACKGROUND

Several essential oils, including citronella (lemongrass, Cymbopogon sp., Poaceae), are well‐known mosquito repellents. A drawback of such products is their limited protection time resulting from the high volatility of their active components. In particular, citronella oil protects for <2 h, although formulations with fixatives can increase this time.

RESULTS

We synthesized hydroxylated cyclic acetals of citronellal, the main component of citronella, to obtain derivatives with lower volatility and weaker odour. The crude mixture of isomers obtained in the reaction was tested under laboratory conditions for its repellency against two mosquito species, the major malaria vector Anopheles gambiae and the arbovirus vector Aedes albopictus, and found to be endowed with longer protection time with respect to DEET (N,N‐diethyl‐meta‐toluamide) at the same concentration. Formulated products were tested in a latin square human field trial, in an area at a high density of A. albopictus for 8 h from the application. We found that the performance of the citronellal derivatives mixture is comparable (95% protection for ≤3.5 h) with those of the most widespread synthetic repellents DEET and Icaridin, tested at a four‐fold higher doses.

CONCLUSIONS

Modifying the hydrophilicity and volatility of natural repellents is a valuable strategy to design insect repellents with a long‐lasting effect. © 2022 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

The mixture of hydroxylated cyclic acetals, obtained from condensation of citronellal with glycerol, was tested for its repellency against mosquitoes and showed a longer‐lasting protection in comparison to DEET.

1. INTRODUCTION

The control of malaria, dengue, leishmaniasis and other vector‐borne diseases can be efficiently carried out by reducing contact with the vector and consequent blood‐feeding. ¹ Topical repellents, house screening, insecticide‐treated bed nets or dog collars currently are used with success for this purpose. Although indoor protection can be obtained with different approaches, such as house screens, bed nets and insecticide spraying (mainly pyrethroids), topic repellents, applied on the skin or clothes, remain the main personal protection method outdoors. There is growing evidence that a substantial part of residual malaria transmission occurs outdoors thanks to behaviourally resistant mosquitoes that previously have been inside houses. ² Moreover most of the vectors of dengue and Chikungunya are daytime feeders, biting outdoors.

Therefore, topical repellents are particularly important to reduce outdoor transmission in areas where vector‐borne diseases are endemic or epidemic.

Several plant extracts are well‐known natural repellents against mosquitoes and other haematophagous arthropods. ³ Volatile compounds are detected through the insect's olfactory system and trigger avoidance behaviour, being perceived as potentially toxic. ³ , ⁴ The main drawback of most plant‐derived repellents is their short protection time, related to their relatively high volatility. An exception is menthane‐3,8 diol (PMD), a compound obtained from the steam distillation of lemon eucalyptus (Corymbia citriodora). ³ The longer‐lasting repellent effect of PMD probably is due to its lower vapour pressure, which in turn is related to the presence of two hydroxy groups in the molecule.

Citronellal is the main component in the essential oils of the citronella grass, Cymbopogon nardus and Cymbopogon winterianus (Poaceae). ⁵ , ⁶ These extracts, commonly indicated as citronella, are the most common natural compounds used as topic and spatial mosquito repellents in commercial products. ³ Formulations of topical repellents contain ≤5–10%, as higher concentrations can cause skin sensitivity. ³ Higher concentrations have been used in citronella‐treated armbands ⁷ which are commercially available. The short protection time of the oil (≈2 h) can be increased by microencapsulation or addition of components, such as vanilline, which do not have a repellent effect on their own. ³ , ⁴ , ⁸

Several additives, such as liquid paraffin, salicylic acid, mustard and coconut oils, have been used in formulations of citronella‐based repellents to increase their protection time. ⁹ Among them, vanillin is the most widely used additive and its effects have been tested with the main dengue vectors, Aedes aegypti and A. albopictus. ⁸ , ¹⁰ , ¹¹ , ¹² however, as a consequence of the limited duration of repellency, the U. S. Environmental Protection Agency requires that labels of citronella‐based insect repellents carry the recommendation to repeat applications every hour. ¹³

Crude citronella oil has been tested mostly against health‐threatening hematophagous dipterans, ⁴ , ⁷ , ¹⁴ but also against other blood‐feeding arthropods such as triatomine bugs Trypanosoma cruzi 15, 16 ticks ¹⁷ and insects attacking stored foods. ¹⁸

To the best of our knowledge, only a few studies have used pure citronellal in such applications. The repellency effect against A. aegypti of 1 mL pure citronellal applied to the forearm lasted <1 h, ⁴ whereas a study testing wristbands impregnated with a 30% solution of citronellal reported a repellency of 78% against the vector of lymphatic filariasis and epidemic encephalitis, Culex pipiens pallens, but did not report its duration. ⁷ Regarding toxicity, citronellal was found to interfere with cytochrome P450‐mediated oxidation. ¹⁹ Moreover, cytochrome P450 and glutathione S‐transferase were mostly inhibited by citronellal, as well as by other monoterpenes present in essential oils, when supplemented to 4th instar A. aegypti larvae together with piperonyl butoxide, a synergist component of pesticide formulations. ¹⁹

As most of the natural repellents are efficient at high concentrations, a strategy to extend their protection time involves the use of derivatives with reduced volatility as reported by Iovinella and coworkers. ²⁰ In this work, three cyclic ketals of menthone, which offer reduced volatility with respect to the parent compound, were found to be active for longer times with respect to DEET against mosquito bites. Another advantage of these derivatives, again as a consequence of their low volatility, was a much weaker odour with respect to both the parent menthone and to other products used as mosquito repellents. Moreover, different derivatives of menthone, including its glyceryl acetal, have been reported to be endowed with insecticidal activity against different species of mosquitoes, such as Culex quinquefasciatus, A. aegypti and Anopheles tessellatus. ²¹

With regards to citronellal derivatives, a patent has reported citronellal acetals against slugs, millipedes and earthworms. ²² More recently, cyclic acetals obtained by reaction of citronellal with ethylene glycol and glycerol have been reported as mosquito repellents. ²³ Such derivatives also act as repellents against the pharaoh ant (Monomorium pharonis), the best being those obtained with ethylene glycol and propanediol. ²⁴

With the aim to produce mosquito repellents with longer protection times, we have synthesized cyclic acetals by the reaction of citronellal with glycerol, and tested their repellent activity against mosquitoes using the human bait method, along with the World Health Organization guidelines. ²⁵ The mixture obtained from the synthesis, contained both hydroxy dioxanes and hydroxymethyl dioxolanes, and was tested against A. albopictus and Anopheles gambiae. Our product was more efficient than DEET and Icaridin, and its protection time was further increased by the addition of vanillin.

2. MATERIALS AND METHODS

2.1. Synthetic methods

All reagents and solvents were from Sigma‐Aldrich (St Louis, MO, USA). DEET, Icaridin and Lagoon Protection® were supplied by Istituto Biochimico s.r.l. (Padua, Italy). A 1‐L, round bottom flask was charged with p‐toluenesulfonic acid (1.0 g) and methanol (200 mL). While cooling in a ice bath, rac‐citronellal (154 g, 180 mL, 1.0 m) was added portionwise (≈10 mL per portion) to the magnetically stirred solution of the acid, so as to maintain the internal temperature in the flask below 10 °C. Trimethyl orthoformate (106 g, 109 mL, 1.0 m) was added to the resulting clear solution and the cooling bath was removed. After the mixture had warmed to room temperature (RT), glycerol (100 mL, 126 g, 1.35 m) was added in one portion and the mixture was concentrated under reduced pressure (≈8 mbar) while heating externally at 60 °C with a water bath. Sodium methoxide (5 g) was added to the residue in the flask and heating was prolonged for 15 min. After cooling to RT, the mixture was partitioned between water (50 mL) and n‐hexane (300 mL). The organic layer was dried over anhydrous sodium carbonate and then evaporated under reduced pressure to give the mixture of acetal products as a pale‐yellow, clear oil (220 g, 96.5% yield).

2.2. Chemical analysis

The reaction products were analyzed by gas chromatography coupled to mass spectrometry on a GC–MS 7820 GC system‐5977B MSD (single quadrupole; Agilent Technologies, Santa Clara, CA, USA), by injecting 1 μL of a 200 ng μL^–1 solution. The separation was carried out using a 19091S‐433UI column (stationary phase, 95% PDMS, 5% benzene; 30 m × 0.25 mm; Agilent Technologies), using helium as carrier gas (1 mL min⁻¹). The oven temperature was programmed as follows: 45 °C (2 min); 10 °C min⁻¹ up to 200 °C (3 min); 15 °C min⁻¹ up to 300 °C (2 min). The injector port was set at 250 °C. Electronic ionization was carried out at 70 V and acquired m/z ranged from 50 to 550. Data were analyzed using masshunter qualitative analysis B.07.00 software (Agilent) and spectra were checked for diagnostic ions expected based on the product structures. Nuclear magnetic resonance (NMR) spectra were recorded in CDCl₃ at room temperature with a Advance DRX 400 spectrometer (401.36 MHz for ¹H and 100.93 MHz for ¹³C; Bruker, Billerica, MA, USA). The proton data are summarized as follows: s, singlet; d, doublet; t, triplet; q, quartet; quint, quintet; dd, doublet of doublets; dt, doublet of triplets; td, triplet of doublets; m, multiplet, and br, broad signal). For referencing the chemical shift scale (δ), the resonances of the not deuterated residual solvent (¹H) or the deuterated solvent (¹³C) were set to the recommended values. ²⁶

2.3. Estimation of effective dose

The repellencies of DEET, Icaridin and citronellal derivatives were evaluated using the human‐bait technique (to simulate the condition of human skin on which repellents will be applied). ²⁷ , ²⁸ , ²⁹ , ³⁰ Aedes albopictus were reared and tested at 26 ± 2 °C, ≥60 ± 10% relative humidity (RH) and a 14 h:10 h, light:dark photoperiod, within Plexiglas cylindrical laboratory cages (35 cm diameter, 60 cm length) with one end closed by a net. During the tests, cages contained ≈150 nulliparous, 4–7 days old, nonblood‐fed females. Informed consent was obtained from four volunteers before they took part in this study. On the day of the bioassay, volunteers had no contact with lotions, perfumes, oils or perfumed soaps. They wore latex surgical gloves, in which a dorsal square area of 30 cm² was cut open. Mosquito‐exposed skin of one hand was treated with 100 μL of ethanol, as negative control. The other hand was treated with 100 μL of the reaction product in ethanol solution at increasing concentrations (0.005%; 0.02%; 0.05%; 0.10%; 0.50%; 1.00% corresponding respectively to 0.17; 0.67; 1.67; 3.33; 16.67 and 33.3 μg cm^–2). The control hand was exposed in the cage before the treated hand using the same test cage. The number of probing host‐seeking females in a 3‐min exposure period was recorded. Before starting each replication, the mosquitoes' propensity to bite was assessed by inserting the control forearm into the cage and trials were continued only if ≥30 females performed probing behaviour within 1 min. Otherwise, the experiment was interrupted and resumed on a different day. The percentage of repellency obtained from all replicates (expressed as percentage protective efficacy, PE%) was calculated, according to the WHO guidelines for efficacy testing of mosquito repellents for human skin, ²⁵ at each dosage using the formula:

$$\mathrm{PE}\%=\begin{array}{l}\left[\left(\text{number probing untreated hand}–\text{number probing treated hand}\right)/\text{number probing untreated hand}\right]\times 100\end{array}$$

2.4. Estimation of protection time

In order to evaluate the protection time of the mixture of citronellal derivatives against A. albopictus, the PE% was measured, under the laboratory conditions described above, at intervals of 1 h throughout 8 h. 100 μL of a 5% ethanol solution, corresponding to 0.17 mg cm^–2 of exposed skin was applied on the test hand of eight volunteers (four females, four males). The control hand was treated with 100 μL of ethanol, as negative control. The protection time of DEET and Icaridin was measured at the same concentration (5%) and in the same conditions. To evaluate the role of vanillin as enhancer of protection time, the PE% against A. albopictus was measured, under the same conditions as above, by applying on the skin of four volunteers 100 μL of an ethanol solution containing 5% of citronellal derivatives (0.17 mg cm^–2) and 1% of vanillin (34 μg cm^–2).

For tests against A. gambiae 100 5–7 days post‐emergence host‐seeking females (reared, maintained and tested at 27 ± 2 °C, ≥80 ± 10% RH and a 12 h:12 h, light:dark photoperiod) were placed in a laboratory metal cage with a net of polyester mesh (36 cm wide × 36 cm deep × 34 cm high); the PE% was measured with 1 mL of a 5% solution of citronellal derivatives in ethanol applied on 600 cm² (corresponding to 83.33 μg cm^–2) of exposed forearm skin of five volunteers. For each volunteer, the test was performed every hour, for 8 h from the application. During each test, the control forearm was inserted in the cage for 30 s to verify that the number of landings and/or probings mosquitoes was ≥10 per 30 s. If <10 females attempted to bite the untreated hand in 30 s, the test trial was discarded and repeated with a new mosquito cage or postponed to the next day.

2.5. Field trials

The protection time of citronellal derivatives was measured in the field against A. albopictus at the concentration of 5% in ethanol. In these experiments, the PE% was measured by applying 1 mL of each product on a surface area of 600 cm² (corresponding to 83.33 μg cm^–2) of the leg (knee to ankle). Protection time of Lagoon® (20% DEET, 0.5% geraniol, 79.5% coformulants) and Icaridin (20% of active compound, polyethylene glycol (PEG) 400 20%, denatured alcohol 29%, Parfum 0.5%, in deionized water) were measured as a positive controls. As a negative control the same PEG formulation was used. A completely randomized design and blind test were adopted. Four volunteers (three males, one female) and four collection sites were selected to match the number of compounds to be tested plus the control (PEG‐based formulation) as suggested by WHO guidelines. ²⁵ The experiment lasted 7 h and was repeated for 4 days. Each day volunteers were treated with a different compound at the beginning of the test; volunteers moved among the four collection sites that were separated by >50–100 m. Rotations were repeated at 15 min intervals and volunteers returned to the first site at 1 h intervals. The number of mosquitoes landing and/or probing on the skin was counted for 5 min at each site. The experiment was performed in the Botanical Garden of the University ‘Sapienza di Roma’ (41° 54′ 12.6″ N and 12° 30′ 59.7″ E) between September and October 2021.

2.6. Statistical analysis

Following Costantini and coworkers, ³¹ we estimated the protection time of repellents by fitting a probit model under the assumption that the protection (p) of the treated skin from a mosquito could be expressed as p = 1 – (T/C) = (C–T)/C where T and C are (respectively) the number of mosquitoes collected from the volunteer exposing the treated skin and that exposing the untreated skin. In the probit model, we considered the total number of mosquitoes collected every hour pooled over all replicates according to treatment and position. The complete protection time (CPT) for a given treatment was estimated from the time elapsed up to the first mosquito probing in each replicate. The median CPT and its confidence interval were estimated using the Kaplan–Meier survivor function procedure.

3. RESULTS AND DISCUSSION

3.1. Synthesis

The acid‐catalyzed condensation between citronellal and various glycols, performed under azeotropic removal of water, ³² afforded samples contaminated with by‐products from the acid‐catalyzed cyclization of the unsaturated aldehyde. To prevent this problem, rac‐citronellal initially was converted into the corresponding methyl acetal, followed by trans‐acetalization of the latter with glycerol. After removal of the volatiles under reduced pressure, this two‐step procedure afforded acetal in high yield and good GC‐MS purity (96.5% and 94.3%, respectively), thereby allowing the use of the crude product without further purification. As expected, ^{33 13}C NMR analysis of the sample revealed the presence of six distinct acetal compounds, whose C2 carbon atoms were found to resonate in the δ_C = 101–105 ppm region. Based on previous spectroscopic studies on glycerol acetals of long‐chain aldehydes, ³⁴ , ³⁵ the NMR constants recorded for the sample were consistent with a mixture of trans and cis 1,3‐dioxolane and 1,3‐dioxane derivatives (Fig. 1), whose relative content was roughly estimated to be ≈3.0:1.9:1.4:1.0 by least‐square fitting of the intensities of selected ¹³C NMR resonances (Supporting Information Fig. S1). In this regard, on the one hand, it should be pointed out that for the six‐membered ring derivatives, switching between the opposite configurations at the methyl‐bearing stereogenic centre in the side chain (C2’) leads to enantiomeric structures, whose corresponding nuclei are isochronous and therefore undistinguishable under the achiral measurement conditions adopted herein. On the other hand, because of the chirality of the 2,4‐disubstituted‐1,3‐dioxolane the five‐membered ring compounds that differ only in the configuration at C2’ bear a diastereomeric (epimeric) relationship to each other. Under these conditions, the corresponding nuclei in each of the pairs of epimers are chemically nonequivalent and may lead to distinct sets of NMR signals. This is what it was observed with the sample under examination here, in particular in the region of oxygenated C atoms (Fig. S1) where, unlike the 1,3‐dioxanes, all of the resonances assigned to the cis‐ and trans‐1,3‐dioxolanes show up as pairs of narrowly spaced lines.

Figure 1.

Citronellal derivatives. Structures and diagnostic ¹H NMR and ¹³C NMR resonances of the diastereomeric 1,3‐dioxane (top) and 1,3‐dioxolane acetals (bottom) from the condensation of rac‐citronellal and glycerol.

Not surprisingly, ³⁶ GC–MS analyses of the reaction product showed only four peaks. These tentatively were assigned to the cis and trans diastereomers of the 1,3‐dioxane and of the 1,3‐dioxolane, respectively, on the assumption that, as much as for the enantiomers of the former, the silicone GC colum employed in our measurements could not separate the epimeric pairs of the latter to an appreciable degree. For all of them the molecular ion at m/z 228 was clearly visible. However, the mass spectra did not allow to assign the peaks to dioxanes or dioxolanes because in all of the spectra the ion at [M‐31]⁺, expected for dioxolanes, was very faint.

3.2. Repellency in laboratory trials

The citronellal derivatives were tested for repellency against the tiger mosquito A. albopictus using the human‐bait test, as described in Section 2.3. Figure 2 reports the average (±SE) of the protection efficacy obtained with four volunteers as a function of the dosage applied compared to those obtained for DEET and Icaridin. The compound reached 95% repellency when used at the concentration of 0.1%, corresponding to 3.33 μg/cm² of skin. Raw data are reported in Table S1.

Figure 2.

Estimation of effective dose in laboratory trial. Protection efficacy (PE%) average and SE against A. albopictus of the mixture of citronellal derivatives, DEET and Icaridin applied on skin of four volunteers at increasing dosages.

In order to evaluate the persistence on the skin of the mixture of citronellal derivatives, we measured its protection time against A. albopictus and A. gambiae (see Section 2.4). For A. albopictus we used a dosage of 0.17 mg cm^–2 of exposed skin (30 cm²) to eight volunteers. DEET and Icaridin also were evaluated under the same conditions. Volunteers were asked to repeat the test at intervals of 1 h throughout 8 h. The results obtained are reported in Fig. 3.

Figure 3.

Estimated protection time in laboratory trials. (A) Protection time against A. albopictus of the mixture of citronellal derivatives (blue curve and dots), DEET (red curves and dots) and Icaridin (green curve and triangles), applied on the skin (30 cm²) of eight volunteers at 0.17 mg cm^–2; protection time against A. albopictus of the mixture of 5% (0.17 mg cm^–2) of citronellal derivatives with 1% (34 μg cm^–2) of vanillin applied on the skin of four volunteers (light blue curve and triangles). (B) Protection time of the mixture of citronellal derivatives against A. gambiae females applied on the skin (600 cm²) of five volunteers at 83 μg cm^–2. Average and SE are reported for each PE% value.

The 5% solution of crude citronellal derivatives showed a protection of 100% for 2 h, and the repellency value always was >95% up to 8 h (Fig. 3(A)). Icaridin had a similar performance, displaying protection of 100% for 3 h and a repellency always >90%. Instead, DEET, used at a concentration four‐fold lower than the one used in most commercial formulations (i.e. 5%), can protect only for 1 h and its efficacy decreases very fast (after 2 h it drops to <90%). The addition of 1% of vanillin to the 5% solution of citronellal derivatives, resulted in an increase of the protection time. In fact, the mixture of citronellal derivatives showed protection of 100% for 5 h and the repellency was >98% for the following 3 h, as reported in Fig. 3(A). This widely used additive with essential oils, commonly employed at higher concentrations, works as fixative by reducing the release rate of the volatile oil resulting in an improvement of protection time. ³ , ⁸

A similar experiment using a mixture of citronellal derivatives without additives was conducted on A. gambiae (Section 2.4) by applying the repellent solution on a larger skin surface (600 cm²). As shown in Fig. 3(B), in this case the product offered protection for ≥6 h. Raw data are reported in Table S2.

3.3. Field trials

In field experiments (Fig. 4), protection from mosquitoes was estimated to be >99% at the beginning for all tested compounds (Table 1): DEET (Lagoon: 20% DEET, 0.5% geraniol, 79.5% coformulants), 100%; Icaridin (20% of active compound), 99.6% (95%CI: 98.9–99.9%); and the 5% mixture of citronellal derivatives 99.0% (95%CI: 98.2–99.5%). Repellency was still >90% after 3.5 h (DEET 99%, Icaridin 98.2% and citronellal derivatives 94.2%), and at 7 h ≈90% for Icaridin (93.8%) and ≈80% for DEET and the mixture of citronellal derivatives (DEET 83.3% and the mixture of citronellal derivatives 79%).

Figure 4.

Estimated protection time during the field trial. The x‐axis indicates the minutes after application and the y‐axis the estimated protection, computed as the percentage of repelled mosquito. Dots represent observed value from four test volunteers, the solid lines the estimated mean protection and the dashed areas their estimated 95% CI. (A) Mixture of citronellal derivatives at 5% concentration, (B) DEET at 20% concentration and (C) Icaridin at 20% concentration.

Table 1.

Results of probit model assessing protection, defined as 1 – (number of mosquitoes landing on treated skin/ those landing on controls) measured during the randomized design and blind field trial test with four volunteers in a temperate area with high presence of A. albopictus. Estimates reported are predicted probabilities at three time points.

Time from start of trial	Mean protection (95% CI)
	Citronellal derivatives (5%)	DEET (20%)	Icaridin (20%)
Start	99% (98.2–99.5%)	100% (99.9–100%)	99.6 (98.9–99.9%)
Half‐time (3.5 h from start)	94.2 (92.9–95.3%)	99% (98.4–99.5%)	98.2 (97.4–98.8%)
End (7 h from start)	79% (76.1–81.6%)	83.3 (80.5–85.8%)	93.8 (92–95.3%)

Despite a lower average protection (Fig. 5) of the citronellal derivatives compared to DEET and Icaridin (Table 1), the rate of effectiveness reduction was comparable to Icaridin (see interaction term Time*Icaridin, in Table S3 and Fig. S2) and slightly better than DEET (see interaction term Time*DEET, in Table S3).

Figure 5.

Estimated complete protection time during the field trial. The x‐axis indicates the minutes after application and the y‐axis the estimated complete protection (i.e. the protection from the first landing/probing detected by operators). Solid lines represent the estimated mean protection for the blank control, Citronellal derivatives at 5% concentration, DEET at 20% concentration, Icaridin at 20% concentration. Dashed areas represent the estimated 95% CI of the mean.

Of the three tested repellents, Icaridin was the best in terms of complete protection time (time to first landing/probing) of 128 min, followed by the citronella derivatives (88 min) and DEET (86 min). For reference, the first probing on untreated skin was observed in the first minute of exposure. The median complete protection time for the 5% solution of citronellal derivatives was 212 min (95% CI 120–320); thus, probing was observed in half of the exposure event at this point, resulting lower than the other two repellents, tested at four‐fold concentration (DEET: 324.5 (317–390), Icaridin: 260 (186–undefined)) (Fig.5).

4. CONCLUSIONS

Under laboratory conditions, the hydroxylated cyclic acetals mixture, as obtained from the condensation between glycerol and citronellal give 99% protection efficacy against A. albopictus when applied on the skin at a dose of 16.67 μg cm^–2.

The citronellal derivatives maintained an efficacy against A. albopictus >90% for 8 h at 0.17 mg cm^–2, much longer than citronellal, reported to protect for <1 h from A. aegypti. ⁴ Under the same conditions, the effect of DEET dropped to <90% after 2 h. When tested on the major malaria vector A. gambiae at a lower dosage (83 μg cm^–2), a 90% efficacy of citronellal derivatives lasted 7 h (PE = 100% for 6 h). The protection time of our products can be increased further using fixative components such as vanillin, a widely used additive known to potentiate the repellent effects of some essential oils. ⁸ , ³⁷ In fact, an ethanolic solution containing 5% of citronellal derivatives with 1% of vanillin showed a protection efficacy >98% for 8 h against A. albopictus. An additional advantage of our citronellal acetals is their much weaker odour as compared to citronellal, which was judged to be quite pleasant by all volunteers involved in the study.

Because of their partial hydrosolubility, due to the presence of a hydroxyl group, the compounds could be easily formulated in hydroalcoholic solutions.

When tested in an area with high presence of Ae. albopictus using a latin square monitoring scheme, a 5% solution of citronellal derivatives remained active at 95% protection for 3 h and at 85% protection for 6 h. These values are lower than those measured with solutions of DEET and Icaridin, which however were tested at a four‐fold higher dose.

Our results therefore indicate, as suggested previously, ²⁰ that the protection time of natural repellents can be improved by reducing their volatility and/or increasing their hydrophilicity, thus paving the way to new long‐lasting insect repellents.

CONFLICT OF INTEREST DECLARATION

The authors declare that they have no commercial or financial relationships that could be construed as a potential conflict of interest.

ETHICAL STATEMENT

Volunteers agreed to take part in the experiments with an informed consent. This study has been approved by the Regional Ethics Committee for Clinical Trials of Tuscany Region with the registered number 20383_spe.

AUTHOR CONTRIBUTIONS

II and FRD conceived the study; II, BC, PC and FRD participated in the design of the experiments and the interpretation of the results; II, BC, PC, AM and FRD performed the experiments; AM and MM made substantial contributions to acquisition, analysis and interpretation of data; and FRD and II wrote the first draft of the manuscript. All authors read, corrected and approved the manuscript.

Supporting information

Figure S1. C‐O region of the ¹³C NMR spectrum (CDCl₃, 21 °C) of the product mixture of 1,3‐dioxolane and 1,3‐dioxane acetals. The assignment of the signals is based on Wedmid et al. (1980), whereas the signals with greyed‐out frequencies indicate minor amounts of by‐products from the synthetic procedure.

Figure S2. Predicted protection as estimated by the generalized linear model based on data from the field trial. The x‐axis shows minutes after the start of the experiment (min exposure) and the y‐axis shows protection expressed as percentage. The three panels represent one product (% concentration) each. The solid lines represent the mean estimated protection, the shaded areas represent the 95% CI of the mean.

Table S1. Effective dose of citronellal derivatives.

Table S2. Protection time of 5% citronellal derivatives.

Table S3. Results of probit model assessing protection (defined as 1 – number of mosquitoes landing on treated skin:those landing on the controls) measured during the randomized design and blind field trial test with four volunteers in a temperate area with high presence of Aedes albopictus. Estimates reported at link scale (probit). SE of parameter estimate shown; z‐value, estimate of SE ratio; Pr(>|z|), statistics for z‐value.

DATA AVAILABILITY STATEMENT

All the data are reported in the supplemetary files.